THEORY OF KNOWLEDGE The 1913 Manuscript Edited by Elizabeth Ramsden Eames in collaboration with Kenneth Blackwell !his ...
81 downloads
810 Views
6MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
THEORY OF KNOWLEDGE The 1913 Manuscript Edited by Elizabeth Ramsden Eames in collaboration with Kenneth Blackwell !his important work by Russell appeared for the first time In 1984 as a part of The Collected Papers of Bertrand Russell and represents an important addition to our knowledge of Russell's thought. In this work Russell attempts to flesh out the sketch implicit in The Problems of Philosophy. It was conceived by Russell as his next major project after Principia Mathematica and was meant to provide the epistemological foundations for his work. Russell's subsequent difficulties in presenting his theory of knowledge, brought on by what he considered to be the devastating criticisms of Wittgenstein, led both to his abandonment of this work and to a major transformation in his thought. The result was the theory of logical atomism.
Theory of Knowledge gives us a picture of one of the great minds of the twentieth century at work. It is possible to see the unsolved problems left without disguise or evasion. Historically, it is invaluable to our understanding of both Russell's own thought and his relationship with Wittgenstein. Cover design: Twenty Twenty Cover illustration: Caroline Fish Philosophy
ISBN 0-415-08298-6
11 New Fetter Lane London EC4P 4EE
29 West 35th Street New York NY 10001
II I I
9 780415 082983
B56 1.54 R96 1ab
"I;
Theory of Knowledge The 1913 Manuscript BERTRAND RUSSELL Edited by Elizabeth Ramsden Eames in collaboration with
Kenneth Blackwell
London and New York
First published in 1984 by George Allen & Cnwin Paperback edition first published in 1992 by Routledge 11 New Fetter Lane, London EC4P 4EE Simultaneously published in the CSA and Canada by Routledge a di\'i>ion of Routledge, Chapman and Hail, Inc, 29 West 35th Street. New York, NY 10001 Bertrand Russell's unpublished letters and manuscripts' McMaster Cnh'ersity 1983. Chapters I-If] of Part 1 of Theory olKwrdedge G George Allen & Unwin Ltd. 1956. Chapters 1\'-\'1 of Part l ' The Bertrand Russell Estate 1914, 1915. Editorial matter' Elizabeth Ramsden Eames and Kenneth Blackwell 1983, 1992. Funds to edit this volume were provided by the National Endowment for the Humanities, the Social Sciences and Humanities Council of Canada and lvkMaster Uni versi ty . Printed in Great Britain by T.J. Press (Padstow) Ltd, Padstow, Cornwall All rights reserved . .so part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data Russell, Bertrand Theory of knowledge: the 1913 manuscript II. Eames, Elizabeth Ramsden I. Title III. Blackwell, Kenneth 121
Library of Congress Cataloging in Publication Data Russell, Bertrand, 1872-1970. Theory of knowledge. Bibliography: p. includes index. I. Knowledge, theory of. I. Eames, Elizabeth Ramsden. II. Blackwell, Kenneth. III. Title. 192 (121) 84-2983 BI649.R9! ISBN 0-415-08298-6
Contents
Introduction Acknowledgements Chronology
vii xxxix xliii
THEORY OF KNOWLEDGE PART J. ON THE NATURE OF ACQlJAINTANCE
Chapter Chapter Chapter Chapter
I
II III
IV
Chapter V Chapter VI Chapter VII Chapter Chapter
VII IX
Preliminary Description of Experience Neutral Monism Analysis of Experience Definitions and lv1ethodological Principles in Theory of Knowledge Sensation and Imagination On the Experience of Time On the Acquaintance Involved in Our Knowledge of Relations Acquaintance with Predicates Logical Data
5 IS 33
45 53 64
79 90 97
PART II. ATOMIC PROPOSITIONAL THOlJGHT
Chapter I Chapter II Chapter III Chapter IV Chapter V Chapter VI Chapter VII
The Understanding of Propositions Analysis and Synthesis Various Examples of Understanding Belief, Disbelief, and Doubt Truth and Falsehood Self-Evidence Degrees of Certainty
105 119 129 136
144 156 167
BIBLIOGRAPHICAL INDEX
179
GENERAL INDEX
191
Introduction
I. BACKGROUND
I:N HIS A UTOBIOG RAPHY Bertrand Russell gives a vivid description of his life and work for the first ten years of the century: the long struggles with the writing of Principia Mathematica which occupied him to the exclusion of almost every other intellectual project during those years; the day by day frustration of facing a blank page when he was unable to resolve the contradictions in logic that he had himselffound; the intensity of the effort of thought involved in this activity; and the personal unhappiness which made the work harder to bear. 1 All of this makes it easy for us to share Russell's experience and understand the direction of his thought. But for the second decade of the century the account in the Autobiography is much less satisfactory. Partially this is because there are a number of divergent themes: the continued sharing of his work and friendship with the co-author of Principia Mathematica, A. N. Whitehead, and the end of this close eollaboration; the return to Cambridge, and the friendship and philosophieal interactions with Ludwig Wittgenstein; the love affair with Lady Ottoline Morrell with its rapture and pain, and later affairs with others; the war and the exhausting and ultimately futile opposition to it. All these different stories, and the striking omission of any but a superficial diseussion of his own intellectual development during those years, have made it difficult to obtain a clear view of how changes occurred in his thought between r910 and 1920. The most common view of Russell's philosophy during this period sees it as the rapid sequence from the outright epistemological and Platonic realism of The Problems of Philosophy of 1912, through the phenomenalist eonstructionism of Our Knowledge of the External W'orld of 1914, to the logical atomism of "The Philosophy of Logical Atomism" of 1918. 2 In this view the influence of Wittgenstein, which I RUs"ell 196 7,144-200. Russell commentary has come in several phases: in the first phase, his contemporaries seemed hemused by the changes of his philosophy and established the reputation of ficklefless which has clung to his work. An example of this is C D. Broad's remark that Russell develops a new philosophy every few years (in "Critical and Speculative Philosophv", 1924, 79). More recent commentary has focused on two aspects of Russell's
2
vn
viii
INTRODt'CTION
Russell referred to as coming in "two waves",3 was evident in the last of the three phases mentioned, and a fairly sharp dichotomy was supposed to exist between each of the phases of the decade. It is strange that Russell, who seldom retreated from recounting his own failures or fauirs, should have not reported the fact that he had written a large part of a major work on theory of knowledge which had been intended as his first important philosophical work after Principia Malhemalica and which he was forced to abandon under circumstances which constituted an "event of first-rate importance in my life", as he reported later to Ottoline Morrel1. 4 In fact the existence of the partial book manuscript was not known until Bertrand Russell's papers were catalogued in 1967, prior to their sale, and, at that time, Russell did not respond to inquiries about it.' It is tempting to speculate on the reasons for this silence on Russell's part, but, in any case, we can see that the changes in Russell's thought between 1910 and 1920 may be illuminated in an important way by the book. Since the Bertrand Russell Archives have been acquired by McMaster University in Hamilton, Ontario, Canada, the incomplete manuscript has been available for study, and several commentators have noted it, and have sought to interpret its incompleteness and its position in Russell's life and thought. 6 Editorial investigation has virtually established that the missing first 142 pages of the manuscript were published in a series of six articles in The Monist of 1914 and 1915, that Russell stopped writing at page 350 (the
3 4
S 6
work, his logical work, especially theory of types and theory of descriptions, for which the locus classicus is Principia Malhematica (19IO-13), and his epistemology in which the main intcrest has been in work done by Russell either prior to 1918, his theory of acquaintance, or his logical atomism which is linked with Wittgenstein's work and dated by "The Philosophy of Logical Atomism". An example of the interest in acquaintance is the work of Chisholm, "On the Nature of Acquaintance: A Discussion of Russell's Theory of Knowledge", in Nakhnikian 1974. An example of the interest in logical atomism as formulated in the period immediately after World War I is Urmson, Philosophical Analysis (1956). Some critics who have progressed to an inlerest in the later work of Russell still make the assumption that there is a sharp break between it and the "phenomenalism" of the work prior to The Analysis of Matter (1927). See Maxwell, "The Later Bertrand Russell: Philosophical Revolutionary", in Nakhnikian 1974. See Eames, Bertrand Russell's Theory of Knav.'/edge (1969), for a discussion of the critical views of Russell's philosophy. Russell, My Philosophical Development (1959,112). Russell to Ottoline Morrell, #!, 123, [4 Mar. 1916]. (As in the case of all of Russell's letters to Morrell cited here, the original is in the Humanities Research Center, University of Texas at Austin, and a microfilm copy is in the Bertrand Russell Archives, McMaster University [hereafter abbreviated "RA"]. In citing Ihe dales of these letters, "pmk." means "postmarked".) This letter is extracted and printed in Russell's Autobiography (1968,57), in the appendix to Chapter I ofVolurne 2. Blackwell and Eames, "Russell's UnpUblished Book on Theory of Knowledge" (1975,4). See Clark, The Life of Bertrand Russell (1975,204-6). Also there are short discussions in: McGuinness, "Bertrand Russell and Ludwig Wittgenstein's 'Notes on Logic'" (1972); Blackwell, "Wittgenstein's Impact on Russell's Theory of Belief" (1974); McGuinness,
INTRODUCTION
IX
last page of the manuscript) and did not resume wfltmg, and that Wittgenstein's criticisms of one or more major theses of the book could not be answered by Russell and led him to give up his work on it. These are the results of our study of the book and of the historical documents which are available concerning it. Now that the entire text in published form is available for study, it is possible to see its full significance in the development of Russell's thought. This significance can be evaluated only in the light of the historical context in which the work was planned, undertaken, and, finally, left unfinished. Our study also reveals the probable revision by which, apparently, the present fourth chapter was added to the first draft of the manuscript. It is the object of this Introduction to make that historical context as clear as possible, given the availability of the records of the events of the period. The period under consideration is that between the time that Russell finished his share of the work on Principia Mathematica, in I9IO, and the time that he completed work on the sixth Monist article, in February of I9I5. The sources for our knowledge of this period are several: Russell's own retrospective account of the period, which we know suffers from serious omissions;7 the other writings by Russell of this period, published and unpublished; Russell's correspondence with Ottoline Morrell, which began in late March of 191 I when they became lovers and continued almost daily during the period in question; his correspondence with others with whom he discussed his work: his student Ludwig Wittgenstein,8 his philosophically interested acquaintance Oliver Strachey," his friend and the European editor ofthe M anise P. E. B. Jourdain, ,,> his niece and philosophic
7 . 8 9 0 1
"The GTundgedanke of the Tractatus" (1974); Eames, "Philip E. B. Jourdain and the Open Court Papers" (1975); Pears, "Russell's Theories of Memory" and "Wittgenstein's Treatment of Solipsism in the Tractatus" in his Questions in lhe Philosophy ofMind (1975), "The Relation between Wittgenstein's Picture Theory of Propositions and Russell's Theories of Judgment" (1977), and "Wittgenstein's Picture Theory and Russell's Theory of Knowledge" U979); Sommerville. "Types, Categories and Significance" (1979); Perkins. "Russell's Unpublished Book on Theory of Knowledge" (1979); Eames, "Response to Mr. Perkins" (1979); Griffin, "Russell on the Nature of Logic" (1980); Sommerville, "Wirrgenstein to Russell (July, 1913): 'I Am Very Sorry to Hear ... My Objection Paralyses You'" (1980); Blackwell, "The Early Wittgenstein and rhe Middle Russel],', in Block IQBI; Lackey, "Russell's 1913 Map of the Mind" (I981); Pears, "The Function of Acquaintance in Russell's Philosophy" (1981), and "The Logical Independence of Elementarv Propositions", in Block 19BI; Griffin, "Russell's ,\luIriple-Relarion Theory of Judgment"; 1982). This period is discussed by Russell in the second volume of the AutobIOgraphy \ 1968) and in his 1959. 102-9, 112. Russell's faulty memory of this period is discussed in Blackwell, "Our Knowledge of Our Knowledge" (1973). WlItgenstein 1974. Correspondence between Russell and Oliver Strachey, RA. Correspondence between Russell and Philip Jourdain, RA; also leners of Philip Jourdain to Paul Carus and to Open Court, Open Court papers, Special Collections, Morris Library,
x
INTRODUCTION
protegee Karin Costelloe (later Stephen),I! and the chairman of the Philosophy Department at Harvard and leading neo-realist R. B. Perry; 12 correspondence relating to the publication of three of the Monist articles in the 195 6 collection, Logic and Knowledge, with the editor, Robert Charles Marsh,!3 and the publisher of the volume, George Allen & Unwin;l4 the papers of the Open Court Publishing Company, I, the publishers of the Monist and Our Knowledge 0/ the External World; and the notebooks and papers from Russell's seminar on theory of knowledge given at Harvard in 1914 and kept by students in that seminar, Victor Lenzen and T. S. Eliot.!6 We find in Russell's work and correspondence of the period immediately following Principia Mathematica a kind of vacillation concerning work in what he called "technical philosophy". What he meant by "technical philosophy" was, first of all, the kind of mathematical logic done in that book, and second, the application of the techniques developed there to problems in philosophy, in science, and in common sense, A description of what the programme for such a philosophy should be and what truths it could be expected to yield can be found in two lectures which Russell gave in Paris in March of 19II, In the first lecture he sets forth the position of "analytic realism" and introduces the term "logical atomism",l) He describes his position as a commitment to realism in theory of knowledge and with respect to universals, and to the use oflogical techniques in the analysis of complexes into the simples which are their constituents-the logical atoms which are the result of the analysis. Sense-data and a priori knowledge of universals and logical truths are the basic ingredients of knowledge on this view, In this lecture he recommends a method of piecemeal, precise analysis of limited topics as appropriate to the philosophy he calls "analytic realism", or "logical atomism", In the second lecture, "The Philosophical Importance of Mathematical Logic", he claims for that discipline success in the analysis of infinity and continuity, and asserts that it will provide the
II
12 13 14 t5 16
17
Southern Illinois University at Carbondale. Letters of Karin Costelloe to Russell, RA. Russell's letters to R. B. Perry, Harvard University Archives, with photocopies in RA; Perry's Jetters to RusseJJ are in RA with some carbons at Harvard. Letters of Robert Charles Marsh to Russell, and photocopies of letters of Russell to R. C . •\iarsh in RA, with the originals in the Houghton Library, Harvard University. Correspondence between George Allen & Unwin and Russell in RA. Open Court papers, Special Collections, Morris Library, Southern Illinois University at Carbondale. Victor Lenzen's notebook is in RA. His "A Theory of Judgment", Thesis. PhiL 9C, 16 May 1914,isin the Bancroft Library, University of California, Berkeley (photocopy in RAJ. T. S. Eliot's notes on the seminar on theory of knowledge, and two papers on theory of objects which appear to be drafts of papers for the seminar, are in the T. S. Eliot papers in the Houghton Library of Harvard University. Russell, "Le Realisme analytique" \191 n.
INTRODUCTION
XI
most formal, contentless, abstract, and hence the most fundamental truths on which philosophy can build a sound theory of space, time, and matter. 18 In spite of this commi tment to technical philosophy, Russell expressed in letters to Ottoline Morrell a revulsion toward this kind of philosophy. I ~ He remarked in later autobiographical writing that this attitude followed the long years of concentrated logical work on "the big book", Principia Mathematica. 2U But he also told Ottoline Morrell that he had not forsaken technical philosophy and had begun rereading philosophers of the past as a preparation for his next major work, which he was not yet ready to begin writing. 21 In fact, he may have been influenced by a desire to work at the kind of philosophy which was of interest to Ottoline Morrell, who was a woman of sophisticated artistic, intellectual, and political interests, but uneducated in symbolic logic and mathematics. Russell attributed to her a broadening of his tastes, interests, and sympathies,22 and he longed to share his intellectual life with her; at their lovers' meetings they read Plato and Spinoza, and from their discussions a book emerged which they referred to as "Prisons", dealing with the ability of philosophy to free the mind and spirit from the trammels of the here and now. This book, the text of which is lost, reached the stage of typescript, and part of it was published in modified form by Russell as a journal article. 23 The borderline between philosophy on the one hand, and religion and normative ethics on the other, occupied part of Russell's interest, and can be seen in the final chapter of The Problems of Philosophy. At the same time that Russell was most influenced by Ottoline Morrell, he met, in October of 191I, a young Austrian engineering student who had come to Cambridge to study mathematical logic with him.14 This was Ludwig Wittgenstein, who became an important part of Russell's life and thought over two decades. Devoted as a student, brilliant as a neophyte logician, vulnerable to anguishing depressions and sensitivities, alive with aesthetic and philosophical insights, he provided the delights, fascinations, and alarms associated with erratic genius. After two terms at Cambridge, Wittgenstein became a dose personal friend whom Russell loved "as a son", and a valued philosophical colleague whom Russell dreamed of as his SUccessor.B Wittgenstein's influence on Russell was at least as portentous 18 Russell, "L'Importance philosophique de 1a logistique" (191 I). A translation of this article by P. E. B. Jourdain, "The Philosophical Importance of Mathematical Logic", appeared in the Monist (1913). 19 Russell to Morrell, #286, II Dec. 1913. 20 Russell 1959, 102; 1967, 153 (the latter passage was composed in 1931). 21 Russell to Morrell, #220,15 Oct. 19 1 1. 22 Russell 1967 ,20;THESIS
its
ry ne
:a re of us ay ry
1d be ,at
ill he 'Iy It, n-
be lis
he ht ppIa
19
he ut be n)n
its
121
(ollstituents. This proposition also, I believe, or rather one which is very like it, is true, and it is in any case very important, for many reasons, to discover whether it is true. It is not impJied by our previous proposition, for we might be acquainted with the constituents of a complex without being aware that they were its constituents. The arguments in its favour cannot, therefore, be the simple and easy arguments which persuaded us of the truth of our former proposition. Let us now consider what is to be said for and against our new proposition. There is, to begin with, a reason for doubting whether certainty is attainable on this question. We may have acquaintance with an object, without being acquainted with our own acquaintance. All objects of sense that we do not attend to seem to be instances. And it is logically evident that there must be instances, since otherwise every acquaintance would entail an infinite introspective series, which is absurd. Hence it may happen that we are acquainted with the constituents of a complex without knowing that we are acquainted with them. This fact makes it difficult to produce evidence proving that we are not acquainted with them. It would seem, however, that, when we are acquainted with an object, our acquaintance with it can usually be discovered by an effort of attention-though it must be confessed that, ifit could not be so discovered, it ishard to see how we could ever know that it existed undiscovered. In order to circumvent this psychological problem, and to put our main proposition into a form which can be tested, let us restate it as follows: We may be acquainted with a complex without being able to discover, by an introspective effort, that we are acquainted with the objects which are in fact its constituents. In this form, our principle is more capable of empirical verification than it was before. The case of an experience may serve to illustrate the kind of argument which may be brought in support of this proposition. We decided that experiencing a given object is a complex, not because direct inspection reveals any complexity, but because experiencing has properties which we did not see how to account for on any other hypothesis. But although we found no difficulty in being acquainted with an experience, the most attentive introspection failed to reveal any constituent of an experience except the object. Thus if we were right in our analysis of experience, this instance affords a complete proof of our present thesis. There is a class of arguments in favour of our thesis, derived from spatio-temporal divisibility, which can only be used, if at all, with great caution. It may be argued that, owing to the infinite divisibility of space and time, any sense-datum which has spatial or temporal extension must consist of an infinite number of parts. If so, these parts are necessarily imperceptible, since they must be too small for our senses. It will follow, therefore, that every sense-datum which has spatial or temporal extension must be a complex of an infinite number of constituents with which we are no!
10
20
30
40
122
THEORY OF K:-';O\\LED(iE
acquainted. Such an argument, however, involves an unduly naIve transfer. ence of infinite divisibility from physical space and time to the spaces and time of the senses. In regard to time, we have already seen how physical time «'.n be logically constructed without assuming that a continuous duration actually consists of temporal parts. And in regard to space, it would seem that a similar construction is possible. There seems no reason to assume that, say, a uniform patch of colour occupying a small visual area must be complex; it is quite possible that the infinite divisibility of physical space results from a logical construction out of data which are not infinitely 10 divisible. We shall consider this question at a later stage; for the present it is only introduced to show that any argument, as regards our thesis, which is based upon the infinite divisibility of space and time, requires close scrutiny and is very likely to be invalid. Cases of apparently continuous change, however, although it is possible to account for them on the assumption that they are composed of a large finite number of parts, each internally constant, are much more naturally and easily accounted for on the hypothesis that they are complexes whose constituents we are not aware of. The usual arguments of philosophers who dislike analysis are largely drawn from such examples as the experience of a 2C visible motion. There seems here to be change without any succession of separate states composing the change. Such experiences are only what is to be expected, if we can be acquainted with a complex, such as a motion, without being able to have more than inferential knowledge of its constituents. And the way in which many sensible objects which at first appear simple are found to be complex as soon as they are attended to is much easier to account for on the hvpothesis that at first they had constituents with which we were not acquainted than on any other hypothesis known to me. The analysis of a complex raises different problems according to the nature of the complex. In particular, it makes a difference whether the 30 complex is one which is determinate as soon as its form and its constituents are given, or whether it is one which even then may be one of several. We call a complex "logically possible" when there is a corresponding proposition. Thus for example "A precedes B" and HB precedes A ,. are both logically possible, though at most one of them can be actual. What is relevant to the problem of analysis is whether. with the same constituents and the same form, several complexes are logically possible, not whether several are actual. In general, in a complex of n terms, there are various "positions" in the complex, corresponding to different relations (generally each of them functions of the relating relation) which the constituents have 40 to the complex. A complex may be called "symmetrical" with respect to two of its constituents if they occupy the same position in the complex. Thus in "A and B are similar", A and B occupy the same position. A complex is "unsymmetrical" with respect to two of its constituents if the two occupy
PART II, CHAP. II AKALYSIS AND SYNTHESIS
'f-
ld
Ile
)ll
m
Ile
be ce
Jy is IS
oy
to ite od )se ho fa of to
)fl,
In-
:ar 1er ith
e.
he :he Ilts Xle
)si-
Jih is nts :ler
)Us
tlly lve wo ;in , is
IPY
123
different positions in the complex. An unsymmetrical complex is called "homogeneous" with respect to the unsymmetrical constituents if a logically possible complex results from interchanging them; otherwise it is called "heterogeneous". Thus "A is before B" is unsymmetrical and homogeneous with respect to A and B. But "A is a constituent of a" is unsymmetrical and heterogeneous with respect to A and a. It is only in cases of unsymmetrical homogeneity that a complex is not determined by its form and its constituents. The relating relation in a complex is always heterogeneous to all the other constituents. Let us denote the other constituents by XI, X2, Xa, ••. Yl,YZ,Y3, .,. z)' Z2' Z.3, •.• , where all the x's are homogeneous, and all the y's, and all the z's, but no X is homogeneous with ay or az, and no y with az, and soon. If we denote the relating relation by R, the complex may be denoted by
10
There will be certain permutation-groups among the x's which leave the complex unchanged. E.g., taking only three constituents, we might have the complex the same for all permutations, or the same when X I and X2 are interchanged, but altered if X3 is moved, or the same so long as the cyclic order is the same but changed if it is reversed, or changed by any permutation. Only the first of these cases gives complete symmetry as regards the x's. 20 The complex is not wholly determinate when its form and its constituents are given, unless it is completely symmetrical with respect to each set of homogeneous constituents. The further formal study of this topic belongs to the theory of permutation-groups, and need not concern us here. We can now consider the main question which concerns us in regard to analysis, namely: what sorts of facts must we perceive in order to be able to analyze? In order to make sure that we are concerned with a genuine case of analysis, it is necessary to make sure that, in the case considered, the complex is really given. Perhaps the best cases are sensational: the discovery of constituents of the visual field, for example; or-to take a case where 30 analysis is harder-the discovery of overtones in a note. In regard to such cases as the selection of one object out of the visual field, the word "analysis" is often misused. Mere selective attention, which makes us aware of what is in fact part of a previously given complex, without making us aware of its being a part, is not analysis. Chairs and tables, we are told, originally form part of one visual continuum; but even if this be so, we can become aware of them as separate objects without performing any act that can properly be called analysis. All that is necessary is that the area of attention should be narrowed, not that we should realize, as analysis requires, the relation of the new partial objects to the old total one. 40 A whole which is to be analyzed must presumably be first itself an object
12~
1O
20
THEORY OF
K~O\X'LEDGE
of attention. The "sensible continuum" of psychologists endeavouring to undo the work of thought is itself a late discovery of thought. All primitive consciousness is selective; the "sensible continuum", which rejects nothing, is the product of the psychologist's scientific impartiality, to which, for the moment, all selection is abhorrent. The visual field as a whole, for example, is not given. No doubt the part of it that we attend to is surrounded by other visual objects, but they grow gradually dimmer, and cannot be regarded as forming altogether a single given object. Before we have anything that can be properly called analysis, we must have a complex given to begin with as an object of attention, whose parts, if they are discovered, do not merely become fresh objects of attention, but are recognized as being its parts. Hence if we are to consider analysis of a visual object, we must take an object small enough for the whole of it to be near the centre of our visual field. It will serve to fix our ideas if we take some object as simple as is compatible with the possession of easily distinguishable parts, say a capital T printed in black on a white ground, and ask ourselves what happens when we analyze this object into separate but related constituents. A capital T will ordinarily be seen as one object whose parts are not separately attended to; indeed, if it forms part of a word, the word will usually be the smallest unit of attention, except to some one who is only just learning to read. We can, however, attend to the T alone, and we can easily perceive that it consists of a vertical and a horizontal stroke, the former being below the latter and beginning approximately at its middle point. All this is so easy to see that it is difficult to realize what a complicated process is involved. The fact that each of the two strokes can be further analyzed, of course introduces further complications; but for the present we will ignore this fact, since it seems possible to regard the two strokes as the primary constituents of the T. Let us call the vertical stroke a and the horizontal stroke b, and let the particular T be called)'. Then the two facts that a is part b
b
T 3G
of)' and that b is part of)' must obviously be known if)' is to be analyzed. But here we come upon a difficulty. Do we first reach the judgment "a is part of )''', or do we first become acquainted with the complex "a-partof-),", and thence arrive at the judgment "a is pan of )'''? This raises problems which will be explicitly discussed in connection with self-evident judgments; but our present difficulty can be made plain without much anticipation of what is to be said later. If we first perceive the complex "a-part-of-),", it would seem as though the very process of analysis which we are endeavouring to explain must be performed in order to pass from this to the explicit judgment "a is part of )'''; thus this judgment, which was to
PAR TIl, CHAP. II
o
e c
0' e
., :r IS
n LS
Y :t [t
Ie n
:e )t
11 st Iy ~r
,11 is )f re :y al rt
d. IS
,t -
es nt
:h
ex
;h
LIS
to
A~AL YSIS
AND SYNTHESIS
125
biC merely one step in the process of analysis, will itself involve as much analvsis as it was to have helpiCd to explain. This method of explaining analvsis is therefore impossible. But the view that we begin with the judgment, without first perceiving the complex "a-part-of-y", leaves it inexplicable how we come to know the judgment, which is certainly not obtained by inference from any other judgment. It would seem, therefore, that the complex "a-part-of-y" must be perceived in a differeat way from that in which y was originally perceived, i.e. it must be perceived as a complex, where the interrelated parts are present to consciousness, and not merely discoverable by subsequent attention. 10 There would seem, therefore, to be two kinds of perception of a complex, namely "simple pereeption", which does not involve acquaintance with the parts, and "complex perception", where the complex is seen as a complex of interrelated parts. But if this is the ease, then we need not begin our complex perception with the complex "a-part-of-y", but may begin it at once with y: there will be a way of perceiving y as "a-R-b", where R is the relation between a and b in virtue of which they form the complex y. Let us see whether this suggestion will give a tenable theory of analysis. Taken in an extreme form, the above theory will involve a very great diffieulty, namely: How shall we know that y, the object of simple percep- 20 rion, is identical with a-R-b, the object of complex perception? This difficulty is so great that I think the theory cannot be entertained unless in a form which makes it appear at least possibly soluble. I think the solution must be sought by a further use of the relation of auention. The difference between simple and eomplex perception seems to depend upon the number of objects of attention simultaneously present. We spoke earlier as though attention always had one object, but this appears to be not the case. When we perceive our T as an unanalyzed whole, it is one object of attention; but if another T is placed alongside of it, we shall naturally say that we see "two 1's". In this case, there seem to be two 30 simultaneous objects of attention. And similarly when we have analyzed our T into a vertical and a horizontal stroke, we shall have two objects of attention, namely the two strokes, where before we had only one, namely the 1. We may therefore suggest that complex perception consists in acquaintanCe with a whole combined with attention to its parts, It will be natural to say, conversely, that simple perception of a complex consists in attention to the Whole combined with acquaintance with its parts. If this is the case, the problem of analysis is merely the problem of transferring attention from the whole to the parts. Let us examine this theory. In order to be able to know that the object of complex perception is 40 identical with the object of simple perception, we only now require to be able to know that an object to which we attend at one moment is identical with an object with which, at another moment, we have inattentive acyuaintance. It is certainly the case that we can know this. We have the
126
THEORY OF K:--:OWLEDGE
experience constantly of becoming attentive to an object which we know was already given in sense before we became attentive to it. This is po~ible because we can attend to an experience which is in the immediate past, even if we did not attend to it or its object when it was present; and thus we can come to know that the object of this past experience was identical with the object of our present attention. But for this possibility, we could not know that attention selects some but not all of the objects of acquaintance. There is, therefore, no difficulty, on our present theory, in accounting for our knowledge that the object of complex perception is the same as the object of 10 simple perception. Our present theory seems to accord well with what introspection reveals concerning such an operation as the analyzing of our T. At first, when we attend to it as a whole, our attention has one object before it, namely the T; but it would seem that even then we are acquainted with the two strokes of which it is composed, although we are not attending to them. Afterwards, however, it is the two strokes to which we are attending: we are still acquainted with the T, but not attending to it. Finally, in order to know that the T consists of the two strokes, we require a complex perception in which we attend to the T and to the two strokes simultaneously, or in which we 20 perceive the identity of the T with which we have acquaintance in the complex perception and the T to which we attend in the simple perception. Al! this seems quite in accord with introspection. It is obvious, however, that if complex perception is to account for analysis, it must not consist merely in simultaneous attention to two objects. Nor can we be content with this together with acquaintance with the complex. Complex perception must involve some consciousness of the relatedness of the two objects. In the complex perception of our T, in which we attend to the two strokes, we are acquainted with their spatial relatedness, though we may not be attending to it. But here the very problem we set 30 out to solve meets us again, not one step advanced towards solution. What is consciousness of the relatedness of two terms? It must not be identified with the simple perception of the complex, for it involves attention to the two terms, which simple perception of the complex does not. The question is: Is there any difference between the complex and "the relatedness of the two terms"? Can we distinguish between the T and the two strokes in a certain spatial relation to each other? We can see why we should use different phrases, because one is appropriate when we are attending to the complex, the other when we are attending to the terms. But apart from this subjective difference, is there any difference between the objects denoted by these two 40 phrases? I think we must say that there is no difference whatever between the two objects. When there are two strokes in a certain spatial relation, there is a T; the T consists of these two strokes so related; the T is these two strokes so
PART II, CHAP. II ANALYSIS AND SYNTHESIS ~
e n
n
e v e r ,f
s e
,
,f
,
:
II .t
i1 e e
r
e
e i1 t S
i1 D
S [)
i1
e )
J
127
related. I do not think it is possible to discover a vestige of difference between the two objects. Hence in being aware of the relatedness of the two strokes, we are aware of the T. (When I speak of the "relatedness" I do not mean the abstract relation.) It would seem, therefore, that, throughout the process of analysis, we are acquainted with the complex and with its constituents, and that what changes during the process is only the direction of our attention. (I am not denying that we are sometimes acquainted with a complex without having any acquaintance with its constituents; but obviously so long as this state of things lasts, analysis cannot begin.) At first, our attention is directed to the complex; then it passes to the terms which are 10 constituents; before the analysis is complete it must pass also to the relation which is a constituent. A complete analytic acquaintance with a complex involves acquaintance with the complex together with attention to the terms and the relation which constitute it. But however necessary the above process may be to analysis, it is still not sufficient. We might attend to terms and a relation, and be acquainted with a complex, without the terms and the relation being the constituents of the complex. We want to reach a proposition of the form: "This complex consists of these terms so related." This requires a simultaneous attention to the complex and its constituents. In order to judge: "This T consists of two 20 strokes at right angles", we must attend to the T and the strokes and their relation-at least, so it would seem. Thus so far nothing effective has been done in the way of analyzing analysis. Before proceeding, let us pause to take stock of what has already been said in this argument. Selective attention alone, we decided, is not analysis; what is to be analyzed must be given as one object of attention. To analyze a given complex y, it would seem that we must know such facts as "a is part of y". But since this cannot be an inferred judgment, it must be derived from perception of a complex "a-part-of-y". This requires us to admit a "complex perception" of a complex, as opposed to a "simple perception". These 30 two differ, it would seem, in respect of attention. In both, we are acquainted with the complex, but we only attend to it in simple perception, whereas in complex perception of the complex it is the parts that we attend to. According to the above, our acquaintance with the complex is of the same nature in simple perception and in complex perception. Now complex perception was forced upon us by the instance "a-part-of-y", where the terms and the relation are certainly given. It would seem to follow that, since We can know that a term is part of a complex, all acquaintance with (omplexes must involve acquaintance with their parts to the extent involved in naming them by means of their parts. That is to say, we do not have to 40 deal with a problem of which the proper statement is "how do we analyze y (where y is a given complex)?", but "how do we analyze aRb (where a and Rand b are known by acquaintance)?" It would seem that, when a complex
128
20
THEORY OF K:--;(l\XLEI)(;E
is given, mere attention will enable us to give it a complex name, such as "aR b", though this attention to the constituents is sometimes a very difficult mental and in certain cases is one which seems never to have been performed by human beings. The question "how do we analyze aRb?" may seem almost childish. It may be said that all the real work of analysis is done when a complex name of this kind has been given. As a matter of psychological experience, this is no doubt true: the effort of attention involved in passing from simple to complex perception of a complex may be very great. But this raises no new epistemological problem. There are, however, certain problems remaining. There are purely logical problems, such as: How is the meaning of a complex name such as "aRb" determined when the meanings of the simple constituent names are known? What is meant by "a is part of aRb"? What is meant by "aRb consists of a and Rand b united in the general form of a dual complex"? And if these questions were answered, there would remain the question as to how we come to know the truth of such propositions. These questions, important as they are, I shall not now discuss. Partly they do not belong to theory of knowledge, partly they belong to a later portion of the subject, partly, wherever they belong, I do not know the answer to them. On the subject of synthesis, there is very little to be said. The psychological process of synthesis, when it is possible, consists of attending simultaneously to certain terms and a certain relation until we perceive a complex formed of those terms so related. The epistemological problems raised by this process are the same as those raised by analysis, and do not require separate discussion.
Chapter III Various Examples of Understanding
~ CHAPTER I, we considered only one particularly easy case of understanding a proposition. In this chapter, we have to consider various examples and difficulties. The first example I wish to discuss is the understanding of propositions which have no constituents, i.e. where all the constituents have been replaced by apparent variables, and the pure form alone remains. We may take as the type of such propositions the proposition "something is somehow related to something", i.e. "there are an x and ay and an R such that x has the relation R to y" or "xRy is sometimes true". This is much the same proposition as "there are dual complexes". It will be remembered that, according to our theory of the understanding of propositions, the pure form is always a constituent of the understanding-complex, and is one of the objects with which we must be acquainted in order to understand the proposition. If this be true, then the understanding of the pure form ought to be simpler than that of any proposition which is an example of the form. Since we desired to give the name "form" to genuine objects rather than symbolic fictions, we gave the name to the "fact" "something is somehow related to something". If there is such a thing as acquaintance with forms, as there is good reason to believe that there is, then a form must be a genuine object; on the other hand, such absolutely general "facts" as "something is somehow related to something" have no constituents, are unanalyzable, and must accordingly be called simple. They have therefore all the essential characteristics required of pure forms. Again, the view that a pure form is something which must be understood in order to understand a proposition which has this form, is satisfied by this view of what a form is. To say that we must have acquaintance with "something has some relation to something" if we can understand "this has this relation to that" is by no means a paradox, and seems in fact a truism. \X'hat tends to obscure this truism, is the absence of attention to the form in ordinary thinking, We may say broadly that it is easier to attend to particulars than to universals, and to universals than to logical forms. But many mental facts involve acquaintance with objects to which no attention is given except in more abstract kinds of thought. For example, the kind of thought which consists in understanding an atomic proposition containing particu-
I
129
10
20
30
,
•..........
130
rHEORY OF K:--;O\XLEDGE
,;
.~
lars in some relation involves attention to its substantial constituents and acquaintance with its form, while the kind of thought which we are at present endeavouring to perform involves attenlion to the form, and acquaintance with the characteristic concepts of molecular proposition&, such as are involved in the words "and", "or", "not". But attention to these concepts will not be required until we come to analyze molecular thought. It is for this reason that, although the understanding of "something has some relation to something" is logically simpler than the understanding of (say) "A is before B", it is nevertheless later in the order of psychological 10 development, which appears to be mainly determined by the nature of the objects to which attention is given. But since we are concerned mainly with the work oflogical analysis of cognitive complexes, we need not be alarmed by an inversion of the psychological order, and we must, therefore, regard the understanding of "something has some relation to something" as logically preceding the understanding of any particular proposition asserting a particular case of a dual relation. The case of a pure form is instructive for several reasons. To begin with, although "something has some relation to something" is a proposition, and is true, it is nevertheless simple; hence understanding and believing, in this 20 case; must both be dual relations. The question therefore arises: How do they differ from acquaintance? And what becomes of the opposition of truth and falsehood in such cases? Again, there are more purely logical questions (which, however, we must leave on one side), such as: How can an object be at once simple and a "fact", in the sense in which a "fact" is opposed to a simple particular and is the sort of object whose reality makes a proposition true? Why, if pure forms are simple, is it so obviously inappropriate to give them simple proper names, such as John and Peter? These logical questions can no doubt be answered, but for our purposes the epistemological questions are more pressing. Those that concern belief, however, must be 30 postponed until we have discussed belief. We are left, therefore, with the question: Is there any difference between understanding a pure form, such as "something has some relation to something", and being acquainted with this object? And if not, is this kind of understanding radically different from the understanding of propositions concerning particulars? I do not think there is any difference between understanding and acquaintance in the case of "something has some relation to something". I base this view simply on the fact that I am unable introspectively to discover any difference. In regard to most propositions-i.e. to all such as contain any constants-it is easy to prove that understanding is different from 40 acquaintance with the corresponding fact (if any): Understanding is neutral as regards truth and falsehood, whereas acquaintance with the fact is only possible when there is such a fact, i.e. in the case of truth; and understanding of any proposition other than a pure form cannot be, like acquaintance, a
-~~
,
PART II, CHAP. III VARIOUS EXAMPLES OF UNDERSTAN'DIN'G
131
two-term relation. But both these proofs fail in the case of a pure form, and we are therefore compelled to rely on direct inspection, which, so far as I can discover, reveals no distinction, in this case, between understanding and acquaintance. Whether, in regard to such propositions, there is any distinction between belief and acquaintance, is a question which must be postponed until we reach the discussion of belief. It is, perhaps, better to the name "understanding" than the name "acquaintance" to the awareness of such propositions. In the classification of mental facts, various different considerations may be brought to bear. There is, first, the logical form of the mental fact concerned-whether it is a dual or treble or quadruple ... relation. Then there is the logical character of the objects concerned. Then, when both the form of the mental fact and the logical character of the objects are given, there is the actual relating relation which is a constituent of the fact. In considering these grounds of classification, one would naturally have expected that the form of the mental fact would be the source of the most fundamental divisions. In obedience to this supposition, we began with dual relations, acquaintance, attention, etc.; and when we came to propositional thought, it seemed at first as if the change was due to the fact that here the cognitive relations concerned were multiple. This whole point of view, however, is erroneous. The dassification of mental facts by the logical character of the objects involved turns out to be far more important than their classification by their own logical form. In connection with analysis, we had occasion to consider a multiple relation called "complex perception"; but although multiple, this relation had more affmities with ordinary acquaintance and attention than with propositional thought, just because its objects were the same as those of ordinary acquaintance and attention. So with regard to understanding: if our analysis in Chapter I of this Part bore even a distant resemblance to the right analysis, different understandings have very different logical forms. Understanding of "something has some relation to something" is a dual complex; understanding of "something has the relation R to something" is a treble complex; understanding of "a has the relation R to b" is a quintuple complex. The distinctive thing that groups these understandings together is the fact that they all involve the pure form of dual complexes among their objects. For this reason, the simplest possible cognitive relation to a pure form belongs to propositional thought, rather than to the kind of consciousness which we considered in Part 1. We thus arrive at a hierarchy of cognitive relations, according to the "abstractness" of the most abstract of the objects involved. If this most abstract object is a particular, we have sensation, imagination, or memory; if a universal, we have conception and complex perception; if a logical form, i.e. a fact containing no constants, we have understanding, belief, disbelief, doubt, and probably many other relations; jfthe logical form is molecular,
lO
20
3C
40
132
IO
20
30
40
rHEORY OF ""OWLEDGE
we have also inference and the knowledge of the propositions of logic and their instances. The one cognitive relation which seems to preserve a fairly constant character throughout is attention, though as the object grows more abstract, attention grows progressively more difficult; perhaps attention to an abstract object is only psychologically possible in combination with attention to other more concrete objects, the number of which tends to increase as the abstract object grows more abstract. It is an interesting question whether the degree of abstractness of an object can be precisely defined. As a first step towards such a definition, we may distinguish stages of specification in passing from a given abstract object towards a particular which is, in some sense, an "instance" of it. (A complex containing one or more particulars and no general terms, such as "all" or "some", is to be reckoned as a particular.) Thus a universal is removed only one stage from a particular: there are atomic complexes which are particulars of which it is a constituent. But a logical form, even an atomic form, is not a constituent of the particulars which have that form. Thus it is, in some sense, further removed from the particular than the universal is. Finally, a molecular form is not even the form of any actual particular: no particular, however complex, has the form "this or that", or the form "not-this". The question of the definition of "degrees of abstractness" belongs to logic, and we shall not here pursue it further. We shall not assume that this notion has any precisely known meaning, but shall merely employ it as affording suggestions concerning the order of our inquiry. The importance of the understanding of pure form lies in its relation to the self-evidence of logical truth. For since understanding is here a direct relation of the subject to a single object, the possibility of untruth does not arise, as it does when understanding is a multiple relation. This topic, however, belongs to a later stage: it will be considered again in connection with self-evidence. The understanding of a pure form is, according to our theory, a logically simpler fact than the understanding of a proposition which is an instance of the form; it is, moreover, part of what actually occurs when we understand an instance. Nevertheless, it is obvious that, in psychological fact, the isolated understanding of a pure form is more difficult than the understanding of an instance. The reason of this seems to be that pure forms are fugitive to attention, and that attention to them is not readily caused except by means of instances. This question belongs to psychology, and is only relevant because it might seem, if neglected, to afford an objection to our view that the understanding of a pure form is simpler than the understanding of an instance. It is necessary to realize, once for all, that what is simpler is not necessarily, or even usually, easier than what is more complex. Attention, as a psychical occurrence, is governed by biological considerations: particulars may be good to eat or likely to kill us, and therefore it is
y e
)
1
)
1
e
s
'i
f i
PART II, CHAP. III VARIOUS EXAMPLES OF UNDERSTANDING
133
usdul to pay attention to them; but logical forms are not edible or hostile, and attention to them is not a cause oflongevity. This sufficiently explains why it is only a few eccentric persons, unusually relieved from the struggle for existence, whose attention wanders to such unimportant objects. The next kind of understanding which calls for consideration is the kind which is only one degree less abstract than the understanding of pure form-I mean, the understanding of propositions which assert that there are instances of a given universal. To avoid any difficulties other than those that we wish to discuss at the moment, let us take the case of a symmetrical dual relation, and consider the proposition "something is similar to something". Only two objects are involved in this, namely similarity and the pure form. The understanding of the proposition must, therefore, be logically simpler than the understanding of"A is similar toB". We might express the proposition in the form: "Something has to something a certain relation, namely similarity." Here the pure form occurs explicitly, and all that is added is the words "namely similarity". It would seem, if our general view is correct, that it must be an error to regard our proposition as asserting that there are complexes such as "A is similar to B", since the propositions that deal with complexes are less simple than the proposition which we are examining. What our proposition involves must be a relation of similarity to the form. Thus understanding of such a proposition is a three-term relation of the subject, the form, and the relation. The necessity for the form is, perhaps, rather easier to see in this case than in the case of a proposition such as "A is similar to B". If we try to say simply "there are similarities", it is at once obvious that what is meant is more fully expressed by "something has to something the relation of similarity"; and in these words the form occurs explicitly. It is to be observed that, if it is true that something is similar to something, then there is a single fact expressed by these words, and that towards this fact we may have a dual relation of the nature of an acquaintance. This may be called perceiving the fact. It cannot be identified, in this case, with understanding the proposition, as it could in the case of the pure form, because here the understanding of the proposition is a three-term relation, and the proposition is of a form which does not logically guarantee truth. And for the same reasons, perceiving the fact cannot be identified with believing the proposition. It might, no doubt, be questioned whether such a proposition as "something has the relation R to something" would have any meaning if it were false, i.e. if, in fact, there were no instances of R whatever. This question is not altogether an easy one. The instances that are discoverable of relations which never hold are none of them atomic relations, that is to say, they are not single relations at all, but results of logical combinations of several relations. It would seem that we never become aware of a single relation
10
20
30
40
134
THEORY OF K>:O\v"LEDGE
except by the help of instances; hence even if there are any relations of which there are no instances, we can hardly expect to know that there are. But in saying that there are relations of which there are no instances, what we should naturally suppose that we mean would be "there are propositions of the form xRy for values of R for which such propositions are false whatever x andy may be". But with our definition of "proposition", we cannot have any reason to believe that there is a proposition which has never been understood, and we cannot know that a proposition of the form xRy has ever been understood, if the R is one of which there is no instance, and with 10 which consequently no human being is acquainted. This instance suggests, what is also suggested by many other considerations, that our definition of a proposition is inadequate. It seems plain that "aRb" has "meaning" provided R is the right sort of entity, and that the question whether R is the right sort of entity depends upon its logical character, and not upon the more or less accidental question whether instances of it actually occur. Also, when we say that "aRb" has "meaning", it seems impossible to maintain that we mean that somebody understands it. If it has meaning, it can be understood; but it still has meaning if it happens that no one understands it. Thus it would seem that we must find 20 some non-psychological meaning for the word "proposition". If such a meaning can be found, a given R may enter into propositions of the form "aRb", even if there are no complexes of this form, and no one ever thinks about R. In such a case, the proposition "something has the relation R to something" will have meaning but be false. Thus this kind of proposition does not have the necessary truth that belongs to propositions such as "something has some relation to something." It follows from our general theory that the understanding of either of the two propositions "a has the relation R to something" and "something has the relation R to b" is simpler than the understanding of the proposition "a 30 has the relation R to b". In the case of a symmetrical relation, the above two propositions are identical, but in the case of an unsymmetrical relation they are different. The fact that the understanding of the above two propositions is simpler than that of "aRb" may be brought into relation with the question of "sense". Let us now consider the understanding of propositions involving non-symmetrical relations; and for the sake of illustration let us take the case of time-sequence. The proposition "something is before something" is, I think, identical with the proposition "something is after something". But "a is before something" is different from "a is after something". Thus we may begin by the understanding of these twO 4C propositions. We decided that, in "a -before-b", a has to the complex a relation which is one function of the relation "sequence", and b has a relation which is another function of the same relation. If we call the complex y, we will say
PART II, CHAP. III VARIOUS EXAMPLES OF UNDERSTA)',;DLSG
[l
e of r
e [l
s h
e II r
.t
:i a rJ
J [l
s
e s J
Y
s e s
e S J
s s y
135
"u is earlier in y" and "b is later in y", to express these relations. Then "a is before something" means "a is earlier in some complex"; and "a is after ;,omething" or "something is before a" means "a is later in some complex". In order to understand this proposition, we need acquaintance with a and "carlier" and the general form of such complexes as "a is earlier in y". This is the same as the form of "a is part of y"; it is the form of dual complexes consisting of a simple and a complex (i.e. a relatively simple and a relatively complex). Such complexes, in the language of the preceding chapter, are heterogeneous and unsymmetrical; being heterogeneous, they do not give rise to the difficulties connected with sense. Thus these difficulties have been eliminated from the understanding of "a is before something", but only by introducing the notion of a complex in which a is earlier. The proposition "a is before b" must be interpreted as meaning "there is a complex in which a is earlier and b is later". This involves the word "and", which is one of the words that indicate molecular complexes. We cannot therefore deal, at present, with the understanding of this proposition, which must be postponed until we come to deal with molecular propositional thinking. This result is curious, for the complex "a-before-b" is atomic, and yet the corresponding proposition is not atomic. It is not very easy to believe that such a difference can exist, and perhaps some other theory of "sense" can be found which would avoid such a difference. But for the difficulty of naming, i.e. of knowing when to speak of before and when of after, we could explain the understanding of "a is before b" more simply than was done in the above account. We may say: There are two understanding-complexes consisting of the subject, the form "x and y in a relation", sequence, A and B, and of these two complexes, one is called the understanding of "A is before B", while the other is called the understanding of "B is before A". But the difficulty here is that we cannot tell which is to be called which. In order to know this, we must be able to explain the separate word before, and this can only be done by the help of a sequence-complex. A better theory of how to understand relations with sense could probably be found; meanwhile, the above theory seems at least not logically refutable. Much remains to be said on the subject of the understanding of propositions, but I shall pass on to belief, disbelief, and doubt.
10
20
30
Chapter IV Belief, Disbelief, and Doubt
OFA R, WE have been concerned with non-dualistic attitudes towards objects, and with non-dualistic properties of objects. We now at last reach dualistic attitudes and properties: the dualistic attitude of belief and disbelief in this chapter, and the dualistic property of truth and falsehood in the next. When I speak of "belief", I mean the same kind of fact as is usually called "judgment". I prefer the word "belief", because it has much more definitely the suggestion of a particular dated event which may be studied empirically by psychology. The word "judgment", on the other hand, is generally employed by idealists, and serves to blur the distinction between psychology and logic. Judgment, according to the idealists, is fundamental in logic, and yet is something which could not subsist if there were no minds, though it is independent of this or that mind. We cannot, in our philosophy, give a statement of this view which shall be prima facie tenable. But we may illustrate it by defining a judgment as the fact "that such-andsuch a proposition is believed". I.e. the judgment "A succeeds B" will be the fact \if it is a fact) "There is a subject which believes that A succeedsB". I do not suggest that this is what idealists mean by judgment, but that it is the nearest approach to their view which is formally possible to us. In this form, it presupposes a discussion of belief, and therefore it is to belief that we must now turn our attention. The word "belief" is, in one respect, less suggestive of correct views than the word "judgment". It is obvious that judgment involves the sort of thing that is true or false, and not merely single existents, such as tables and chairs. "Belief", on the contrary, is very often supposed to be possible towards the very same kind of objects as can be given in sensation. "Seeing is believing" embodies a confusion from which many philosophers have suffered. This confusion, for example, underlies Hume's account of Belief, I which begins with the words:
S
10
20
30
The idea of an object is an essential part of the belief of it, but not the whole. We conceive many things, which we do not believe; 1
TreatlSe. Bk.
I,
Part Ill, Section VII.
136
PART II, CHAP. IV BELIEF, DISBELIEF, AND DOUBT
137
froIU which it passes naturally to the well-known conclusion: As belief does nothing but vary the manner, in which we conceive any object, it can only bestow on our ideas an additional force and vivacity. An opinion, therefore, or belief may be most accurately defined, A LIVELY IDEA RELATED TO OR ASSOCIATED WITH A PRESENT IMPRESSION.
This account assumes that the object of a belief is the same as the object of an idea. The confusions which have led to this assumption are many, but as il has had disastrous consequences in philosophy, it is important that a sufficient number of them should be set forth. In the first place, our theory admits that there is a kind of mental fact, called "understanding a proposition", which does not involve the opposition of belief or disbelief, but shares the neutrality of mere acquaintance, and yet consists in a complex of exactly the same form as a belief. When a certain subject understands a proposition and when he believes it, the two complexes differ, according to us, solely in the fact that understanding and believing are different relations: belief in the proposition results from substituting believing for understanding in the complex which is the understanding of that proposition. But understanding a proposition, if our previous theory on this point bore any resemblance to the truth, is a complex of a wholly different form from acquaintance, except in the one case of understanding a pure logical form. Both belief and understanding, except in this one case, have not a single object, the "proposition", but have a plurality of objects, united with the subject in a multiple relation. Thus although there is a neutral attitude, namely understanding, which gives complexes having the same logical form as beliefs, yet this attitude is fundamentally different from that of "having an idea of an object", which, in any sense in which we can admit it as fundamental, must be identified with acquaintance, whether of sensation, memory, or imagination, or (in the case of a universal) of conception. Bume defends himself against the accusation of confusing judgment with conception in a footnote to the above-quoted definition of belief. Be says: \Ve may here take occasion to observe a very remarkable error, which being frequently inculcated in the schools, has become a kind of established maxim, and is universally received by all logicians. This error consists in the vulgar division of the acts of the understanding, into conception, judgment and reasoning, and in the definitions we give of them. Conception is defined to be the simple survey of one or more ideas: Judgment to be the separating or uniting of different ideas: Reasoning to be the separating or uniting of different
10
20
30
40
138
10
THEORY OF K:-';OWLEDGE
ideas by the interposition of others, which show the relation they bear to each other. But these distinctions and definitions are faulty in \'ery considerable articles. Forfirst, 'tis far from being true, that in every judgment, which we form, we unite two different ideas; since in that proposition, God IS, or indeed any other, which regards existence, the idea of existence is no distinct idea, which we unite with that of the object, and which is capable of forming a compound idea by the union .... Whether we consider a single object, or several; whether we dwell on these objects, or run from them to others; and in whatever form or order we survey them, the act of the mind exceeds not a simple conception,
The example of existential propositions, which Hume adduces in the above passage, has no doubt been a potent ally of the theory which identifies judgment and conception. William James, whose neutral monism, we found, involves this same identification, has also, I think, been led to it by the consideration of existential judgments. We seem to judge that objects of sense exist, and to add nothing, in so judging, to what is already given in sense. But the fact is that the whole conception of existence is the result of a confusion between descriptions and true proper names. Of an actually given 20 this, an object of acquaintance, it is meaningless to say that it "exists". But the very same word which, at one moment, is used as a true proper name for a given object, may be used the next moment as a description. We may say "this exists", meaning "the object of my present attention exists", or "the object I am pointing to exists". Here the word "this" has ceased to function as a proper name, and has become a descriptive word, in which an object is described by its properties, and the question may be raised whether there is such an object, since descriptions to which nothing corresponds can be made up. When we say "the King of England exists" we are not uttering a tautology, or adding nothing to "the King of England"; we are saying that 30 there is a person who may be so described. "The King of France does not exist" is true, but would be a contradiction if "exists" added nothing to "the King of France". In a word, "A exists" adds something to "A." if "A" is a description, and is meaningless if "A" is a true proper name. The importance of this result, as regards our present question, is very great. A great many propositions can be thrown into an existential form: "some men are wise" may be put into the form "wise men exist"; "the son of Philip conquered Persia" may be put into the form "the son of Philip who conquered Persia existed", and so on. Those propositions that cannot be put into this form can be regarded as denying existence: "all men are mortal" 40 can be stated as "immortal men do not exist", and so on. So long as it was thought that the "existence" spoken of in such propositions was something which could be significantly predicated of an actual particular, it was
PART II, CHAP. IV BELIEF, DISBELIEF, AND DOUBT
139
impossible to answer Hume's contention that existence adds nothing to the subject, and therefore it was impossible to show how propositions differ from concepts, and how belief differs from acquaintance. Closely connected with this question of existence is the question of the relation of the grammatical subject to the real analysis of a proposition. In "immortal men do not exist", there is not an entity, "immortal men", of which non-existence is predicated. "Immortal men" is merely a grammatical constituent of the phrase, which does not correspond to any actual constituent of the proposition, of which a more correct expression is "whatever is human is not immortal". And this applies to many grammatical subjects which pass muster as complex concepts. In such cases the complexity is usually propositional, i.e. the propositions which grammatically have these subjects are really molecular, and involve a separation of what appeared to be the parts of one complex subject. Thus it is wholly misleading to suppose that, wherever a grammatical subject is formed by the combination of names of concepts, there must be a complex concept formed by combining the concepts whose names occur in the grammatical subject. This misleading supposition has quite unduly extended the general belief in complex concepts among philosophers, and has obscured the part played by relating relations in forming complexes, as well as the fundamental difference between a concept and a proposition. In the course of the discussion on belief, Hume mentions the understanding of propositions which are not believed. He says: Suppose a person present with me, who advances propositions, to which I do not assent, that Caesar dy'd in his bed, that silver is more fusible than lead, or mercury heavier than gold; 'tis evident, that notwi thstanding my inc red ulity, I clearly understand his meaning, and form all the same ideas, which he forms. My imagination is endow'd with the same powers as his; nor is it possible for him to conceive any idea, which I cannot conceive; nor conjoin any, which I cannot conjoin. It might be thought, from this passage alone, that Hume would distinguish between conjoining ideas and conceiving a new complex idea, but this would be a misinterpretation, as appears from the context. But what does appear from the above passage is, that Hume conceives thought as conjoining the ideas of objects, while what makes a thought true is a conjunction of the objects. This gives, of course, a short and easy way of defining falsehood, and of distinguishing between propositions and the facts that make them true. For us, owing to our rejection of "ideas" as a tertium quid between subject and object, no such explanation is possible. When we judge that mercury is heavier than gold, mercury and heavier and gold must themselves
10
20
30
40
140
10
20
THEORY OF Kl"OWLEDGE
be constituents of the event which is our judging: we cannot say that We bring our idea of mercury into some relation with our idea of gold, but that mercury and gold do not themselves stand in the "corresponding" relation. It is curious that authors who believed in ideas were not troubled as to this correspondence of relations. The relation between my idea of mercury and my idea of gold cannot be "heavier", since my ideas are not supposed to have weight. Nor can it be the idea of "heavier", since that is not a relation. It must, therefore, be some new relation, in some way related to "heavier", subsisting between my ideas, but not necessarily present to my consciousness when I judge. This, however, is obviously absurd. My Judging obviously consists in my believing that there is a relation between the actual objects, mercury and gold, not in there being in fact a relation between my ideas of these two objects. Thus the whole nature of belief must necessarily be misunderstood by those who suppose that it consists in a relation between "ideas", rather than in the belief of a relation between objects. Something subjective must be so inseparably bound up with belief as to make it impossible to regard it as a dual relation to a single object, since, if we did so regard it, falsehood would become inexplicable. But the champions of "ideas" introduce the subjective element at the wrong place. It comes in with the subject itself, and is more important than in the case of acquaintance because, since belief is a multiple relation, it does not involve anyone object corresponding to the belief and not involving the subject. But the several objects which are constituents of the belief are just as free from reference to the subject as the one object in the case of acquaintanceexcept, of course, when the belief happens to be about the subject. What William James says about belief in his Psychology is, from our point of view, much the same as what is said by Hume. At the beginning of his discussion 2 he says: Everyone knows the difference between imagining a thing and believing in its existence, between supposing a proposition and acquiescing in its truth. In the case of acquiescence or belief, the object is not only apprehended by the mind, but is held to have reality. Belief is thus the mental state or function of cognizing reality .... In its inner nature, belief, or the sense of reality, is a sort of feeling more allied to the emotions than to anything else.
30
What has been said already as regards existential propositions applies unchanged to the first part of the above quotation. "Cognizing reality", according to our theory, would more aptly describe acquaintance than belief. Is mistaken belief cognizing of unreality? Are there not beliefs which 40
2
Chap.
XXl,
"The Perception of Reality", Vol.
1I,
p.
283.
PART II, CHAP. IV BELIEF, DISBELIEF, AND DOl'BT
141
afe in no wav existential? Is not the whole conception of "existence"
inapplicable to just the objects of which we are most certain, namely those given in present acquaintance? And are not imagined objects also "something", with a "reality" of the kind appropriate to such objects? I shall not again enlarge on these questions, but shall pass at once to the emotion of belief. As regards the "emotion" of belief, the only thing that needs to be understood here is that, however real and important it may be as a psychical fact, it does not concern epistemology, and must be noticed only to avoid the confusions which might result from its unobserved intrusion, like an un desirable alien whose photograph is furnished to the authorities at the frontier. There is an emotion of conviction, capable of many degrees, arising with judgments that hold our unwavering assent, or with perceptions that put an end to a doubt. But this emotion, though it often accompanies judgments, does not by any means constitute them; in fact, it may be exactly the same in the case of two different judgments. A person of a patriotic disposition will iee! exactly the same emotion of conviction in entertaining the belief that his country is the best in the world, as in entertaining the belief that his school or club is the best in the country. The emotion is not a relation to the objects of the belief, but a fresh mental fact, caused, perhaps, by the belief, but quite distinct from it. And it would seem that its intensity is not really proportional to our certainty, but to the energy with which we repel doubt. ::\0 one feels much of this emotion in contemplating the facts in the multiplication-table, because doubt of them is not conceived to be possible. But religious and political beliefs, just because they are denied or doubted by so many, rouse the utmost fervour of conviction. Such beliefs, however, are not those which a philosopher should take as his model. The failure to make a sufficiently radical distinction between belief and presentation has caused difficulties in the theory of erroneous belief, and has led to attempts to introduce the dualism of true and false among objects. There are supposed, by some, to be real and unreal objects, or true and fictitious objects. Such distinctions are always due to the intrusion of belief under various disguises, or at least of the understanding of propositions. One of the most deceptive disguises is the grammatical subject: the golden mountain, or the round square, are thought to be objects, because they can be grammatical subjects; they are therefore declared to be "unreal" objects. All this rests, as already explained, upon a wrong analysis of propositions. The dualism of true and false, with all its attendant distinctions, presupposes propositions, and does not arise so long as we confine ourselves to acquaintance, except, possibly, in the case of abstract logical forms; and even here there is no proper dualism, since falsehood is logically impossible in these cases. But converselv, when we are dealing with belief, we must always bear in mind the duali~m of true and false, and avoid theories based
10
:;'1
20
30
40
142
THtORY OF K?>:O\x Cz, ... C n' The new complex is molecular, and is non-permutative as regards its 20 atomic constituents x1C1a, X2C2a, ... xnCna; also each of these atomic constituents is non-permutative because it is heterogeneous. Whether any difficulties arise from the fact that the molecular complex is still permutative with respect to the constituents of its atomic constituents, is a question which must be left until we come to deal with molecular thought. But it seems fairly evident that no difficulties can arise from this fact. The necessity for the above process is concealed by the fact that it is presupposed in ordinary language. When different complexes can be composed of the same constituents, it is essential that language should distinguish between them. Hence language cannot well express what is prior to 30 these distinctions. If we are to name one out of several complexes composed of the same constituents, we can only do it by means of some such process as the above; hence such a complex always has rather a description than a complex proper name. But since names for unsymmetrical relations always take account of difference of sense, it is very hard to express what can be known before this distinction is taken into account. Having now obtained a non-permutative complex associated with any given complex, we can see how to avoid the difficulties, as regards the theory of truth, which arise from permutativeness. In stating a proposition, there is
l-t8
THEORY OF
K~()\X'LEDGE
always an indication, whether by the order of words, or by inflexions, or in some other way, as to the positions which the objects are to occupy in the "corresponding" complex whose existence is asserted. It follows that What is directly asserted is the non-permutative associated complex. When We assert "A is before B ", we are asserting "there is a complex Y in which A is earlier and B is later." It is impossible to find a complex name which shall name this complex y directly, because no direct name will distinguish it from "B-before-A". Complex names, in fact, are only directly applicable to non-permutative complexes, where the mere enumeration of simple names 10 determines the complex meant. Thus the only propositions that can be directly asserted or believed are non-permutative, and are covered by our original simple definition. Owing to the above construction of associated non-permutative complexes, it is possible to have a belief which is true if there is a certain permutative complex, and is false otherwise; but the permutative complex is not itself the one directly "corresponding" to the belief, but is one whose existence is asserted, by description, in the belief, and is the condition for the existence of the complex which corresponds directly to the belief. In the case we took, if I have a belief whose objects appear verbally to be R, Xl, X2, ... X n , there are really other objects, 20 expressed by inflexions, order of words, etc., and what I am really believing is: "There is a complex y in which XICly, X2C2Y, ... xnCny". In the sense already explained, this proposition is non-permutative, and, except in so far as belief in molecular propositions introduces new complications, our first simple account of the correspondence which constitutes truth applies to it. The actual complex Y itself, whose existence is affirmed by description in our associated molecular complex, cannot be directly named, and does not directly correspond with our belief, or with any possible belief. We may have acquaintance with it, and we may have descriptive knowledge of it; but a complex name for it must be descriptive, not simply composed of the 30 names of the constituents. Belief only reaches it at the second remove, by corresponding with its associated non-permutative complex. It is this that makes permutative complexes so difficult to deal with and such fertile sources of error. Our positive theory of truth has now been enunciated. It remains to consider the arguments for and against it. It must be confessed, to begin with, that its range is as yet very limited. We have seen that, when a complex is permutative, there is no atomic belief corresponding to the complex; and although we have seen how a belief is possible which is not atomic and is true when there is such a complex and 40 false otherwise, yet this belief, because it is not atomic, raises problems which we do not wish to consider until the next Part. Thus the only beliefs with which we can deal at present are those whose truth demands the existence of an atomic non-permutative complex. This includes beliefs in
PART II, CHAP. V TRT.:TH AND FALSEHOOD
149
subject-predicate propositions, and in propositions asserting symmetrical relations or relations which, where they are not symmetrical, are also not homogeneous. We may revert to our former instance, "A and B are similar". The belief in this proposition is true when there is a complex whose constituents are A and B and similar, while otherwise it is false. This is the theory which we have to examine. The chief ground in favour of such a theory must always be that it satisfies our feelings as to what is obvious concerning truth, and that it adds nothing except analysis. Now if we ask: "When is the belief that A and B are similar true?" the only possible answer is: "When A and Bare similar"-that is to 10 say, when there is a complex composed of A and B and similarity. It would seem, therefore, that we have merely expressed what is obvious. This is the whole of the positive argument in favour of our theory. What remains to be done is to show that other proposed theories are inadequate, and that any objections we can discover to our theory can be answered. Theories of truth other than the above may be divided into two classes, according as they do or do not define truth by a correspondence. Of theories which do not define truth by a correspondence, we may distinguish two, as being specially wide-spread, namely (r) the coherence-theory, and (2) pragmatism. Of theories which use a correspondence, there are two sorts, 20 namely those which (as ours does) define falsehood by the absence of a correspondence, and those which define falsehood by correspondence with an object of a different kind from the object which corresponds in the case of truth. Of the first sort, we shall not consider other forms than ours; of the second sort, there are two, namely those which admit real and unreal objects for the correspondence, and those which admit true and false objective "propositions" for the correspondence. We have thus four theories to consider successively. (r). The coherence-theory is generally advocated-for example in Mr. Joachim's Nature of Truth-in connection with a logic wholly different from 30 ours. The chief arguments against it are arguments against the logic with which it is commonly associated; but we shall here assume these arguments, and confine ourselves to a statement and a refutation which assume our logic. A certain relation of "coherence" is supposed given or defined-no matter how-between beliefs, this relation being symmetrical and transitive. It is then supposed that the relation of coherence holds between any two beliefs which belong to the field of the relation-in other words, that anv two beliefs which are coherent with themselves are coherent with each other. Hence the field of coherence forms one group, of which any two 40 members are mutually coherent. This group of beliefs is called true, while other beliefs are called false. The coherence-theory is not stated in the above form by any of its
150
THEORY OF KKO\X'LEDGE
advocates, I have, however, stated nothing beyond what is required to insure that all beliefs shall be divided into two mutually exclusive groups, the true and the false, together with the doctrine, accepted by all advocates of coherence, that any two true propositions are mutually coherent. We may therefore take the theory, in the form above stated, as a sufficient text for criticism. The difficulty of the coherence-theory lies in pointing out any relation, not defined in terms of truth, which holds between every pair of true beliefs and between no pair of false beliefs. The beliefs in dreams may be mutually 10 coherent, in any sense of the word which seems admissible. It is true that they are not coherent with the beliefs of waking life; but that gives no reason why dreams should be rejected rather than waking life. If we knew of one belief which could be known to be true independently of coherence, we could regard the system of true beliefs as the system of beliefs cohering with that one; but if truth is to be defined by coherence, such a course is not open to us. If there is more than one group of coherent propositions, and if all true propositions form one group, then plainly coherence alone does not insure truth. Hence a false proposition must not be coherent even with itself. The coherence-theory in fact involves something very like the doctrine of pre20 Kantian rationalism, that all false propositions are self-contradictory. The only logic, however, which has so far proved adequate to producing this result, has, like the Djyun that fetched water, continued its work too long, and proved in the end that all propositions are self-contradictory. With the application of this result to the special case of the coherence-theory, we may leave this subject and pass on to our next theory of truth. (2). Pragmatism, or at least the pragmatic theory of truth, may be defined as the theory that the truth of a beliefis constituted by some characteristic of its consequences. This way of stating the theory is hardly fair to William James, in whom it was bound up with neutral monism, and with the view 30 that belief does not differ from "entertaining an idea" except emotionally. We have already discussed both these views; but one remark seems appropriate at this point. William James recognized "acquaintance" in the case of sense-perception, but not in the case of memory or imagination. He would not have said, as we should, that these always are acquaintance with something, though perhaps not with anything precisely similar to some sense-datum. Memory, imagination, and conception were, to him, indirect ways of cognizing sense-data: they were true when they were calculated to lead to expected sense-data, and false otherwise. The rest of his pragmatism, I think, grew from this root. But to us, this doctrine is impossible, 40 in virtue of our earlier discussions. We must, therefore, if we are to discuss any form of pragmatism at this stage, discuss a form which recognizes that a belief is a unique kind of mental fact, involving a different sort of relation to objects from any involved in presentation or in "entertaining an idea".
PART II, CHAP. V TRUTH AND FALSEHOOD
151
Let us endeavour to state the pragmatic theory of truth in a form at once as general and as precise as possible. It is not necessary to say that the .:onsequences of the belief must be "good"-we will merely suppose that there is some property u which they have if the belief is true but not if it is false. There is very considerable difficulty as to what is meant by "consequences". It seems fairly plain that logical consequences are not what is meant. The theory boasts of being psychological, and therefore cannot take account of logical consequences which perhaps no one perceives; moreover logic itself must submit to the pragmatic test. Thus the consequences intended must be causal. Now all statements as to the effects of particular events must have a sort of rough-and-ready on-the-whole in-the-long-run kind of character, because in any given case the normal effect may be prevented by something else. We cannot say that a belief is false in a particular instance, because its normal effect has been prevented by some exceptional extraneous circumstance. We shall have to say, therefore, that belief in a certain proposition is trlle when, as a rule, in the long run, its consequences, on the whole, have the property a which we regard as distinctive of truth. Concerning this theory, it is, in the first place, utterly incapable of precision. How many exceptions are allowed by "as a rule"? How great a lapse of time is involved in "the long run"? How great a majority of consequences having the property a is required by "on the whole"? Yet without these qualifying phrases the theory becomes immersed in paradox. Again, how are we to discover the effects of a belief? A few immediate effects may be fairly obvious, as for example when a man faints on hearing himself sentenced to death. But remoter consequences must always be very conjectural, and yet without taking account of them the theory will not work. A third objection is that the theory, contrary to the expressed intention of its advocates, becomes, after all, a very complicated and very arbitrary form of correspondence-theory. A belief is true when there is a certain corresponding fact, namely that its consequences (with the necessary qualifications) have the property a. But why a belief should be called "true" in this case, it is impossible to see. If there were any objection to correspondence-theories in general, it would apply to this theory as to others; but the objections of vagueness, difficulty of discovery, and arbitrariness are peculiar to this theory. Almost the only argument alleged in its favour is, that it is supposed to afford an easy criterion of truth; how false this supposition is, the above brief remarks have, I hope, sufficed to show. (3). I come now to theories of twofold correspondence, i.e. of correspondence of true propositions with one sort of object, and false propositions with another sort of object. And first we will consider the view that true and false propositions correspond respectively with real and unreal objects. This view seems bound up with the existential view of judgment. If
10
20
30
40
152
THEORY OF K::-;O\n,EDGE
e\'ery proposition is of the form ",4 exists", then a proposition is true when its subject is real, and false when its subject is unreal. There must, on this view, be unreal objects, because propositions of the form ",4 exists" certainly are sometimes believed when they are false, yet, on this view, there must be an object A, or else the proposition "A exists" would be meaningless, which it plainly often is noL Hence we can divide objects into real ones, which exist, and false ones, which do not. True propositions are those asserting the existence of real objects, and false propositions are those asserting the existence of unreal ones. 10 There is, to begin with, a trivial objection, which is that the theory, as above stated, forgets negative propositions. In their case, truth implies the unreality of the subject, while untruth implies its reality. This addition is easily made to the theory, but the serious objections remain unaffected by it. The objections are (a) that many propositions are not existential, (b) that when a proposition is existential, its grammatical subject does not represent an actual constituent of the proposition, which, when rightly analyzed, contains no constituent corresponding to the grammatical subject, that there cannot possibly be such things as unreal objects, and that any theory which assumes or implies that there are must be false. All these objections 20 have been explained at length already, and need not detain us now. It is only necessary to observe that what has been said about unreality applies unchanged when it is called by some title of politeness, such as "being for me" or "being for thought", which represent merely the vacillating regret in pronouncing sentence of non-existence on life-long friends. And the same applies to any philosophy which believes, in any ultimate way, in a realm of "possibles" which are not actual. The view that the possible is something, but not quite so much something as the actual, and that error consists in mistaking the possible for the actual,2 is only rendered possible by the wrong analysis of sentences which results from confusing descriptions with 30 proper names. (4). A second form of twofold correspondence theory regards beliefs as having objects of a different kind from the objects of presentations, and divides the objects of belief into two classes, the true and the false. In order to differentiate this theory from its predecessor, we must suppose that propositions (as we will call the objects of belief) are not fictitious when they are false, any more than when they are true: there are such objects in either case. Meinong, whose theory approximates to this type, in the end falls under our previous heading, for he seems to hold that, although "there are" false propositions, yet false propositions do not subsist, whereas true prop40 ositions do subsist. The theory in question has, however, been held; for 2
A view which seems to be that of Professor Stout, "The Object of Thought and Real Being", Proceedings oJ the A rislOlelian SOCIety, 191 C-l1.
PART II, CHAP. V TRUTH Al\D FALSEHOOD
153
example, there is an author who states it in the following terms: "There are., apart from and independently of judgment , true and false propositions, and either kind may be assumed, believed, or disbelieved.'" This theory has an engaging simplicity, and it is hard to find conclusive arguments against it. Nevertheless, I believe it is false. The arguments in favour of the view which are employed in the above article all depend upon the assumption that, when a molecular proposition which is true appears to contain atomic constituents which are false, the apparent atomic constituents must really be constituents. We cannot enter into this question until we come to Part III; for the present, I shall assume by anticipation that a different analysis of such molecular propositions is possible. The arguments against the view we are considering are not of a very definite kind. There is a strong natural conviction that when a judgment is false there is not, in the world of objects, something, pointed to by the judgment, which there is when the judgment is true. Most of the ways of giving effect to this natural conviction are demonstrably false, but the way advocated in the first part of this chapter is, so far as I know, not demonstrably false. Again, it is very difficult to believe that there are objective falsehoods, which would subsist and form part of the universe even if there were no such thing as thought or mind. But the chief objection is that the difference between truth and falsehood, on the theory in question, has to be accepted as ultimate and unanalyzable, whereas it seems obvious that the difference between truth and falsehood must be explicable by reference to fact, i.e. to what is actually in the universe whatever we may see fit to believe. I do not pretend that these arguments are logically compelling; they seem, however, sufficient to make us prefer, if possible, a theory which dispenses with objective falsehoods. And to me, now, it seems obvious, as a matter of inspection, that belief is a mUltiple relation, not a dual relation, so that belief does not involve a single object called a "proposition". But I should not insist upon this argument if it stood alone. There are a variety of objections which may be urged against our account of truth and falsehood. Most of these involve questions which concern molecular propositions, and which therefore must be postponed. There are, however, three kinds of objections which may be briefly considered at this stage. (I) It may be said that the correspondence between belief and fact is arbitrary in our account. ~2) It may be asked how truth and error can ever be distinguished if our account is correct. (3) It may be said that, after all, there must be non-mental "propositions" as opposed to complexes, and that therefore beliefs had better be interpreted as dealing with propositions. (I). Is the correspondence of belief and fact arbitrary in our definition? 3 .. Meinong's Theory of Complexes and Assumptions (III)", Mind, n.s. >-1o. 52, p. 522 .
10
20
30
40
154
THEORY OF KNO\'\'LEDGE
The word "arbitrary" is not a very easily definable word. But the feeling underlying the question may be made definite as follows: The word "truth" has a known meaning, in the sense that, if we understand a belief, we understand also the proposition that the belief is "true". Hence it is not open to us to give any definition we choose of the word "true", and defend ourselves on the ground that definitions are merely verbal conventions. What we call a definition of truth is really an analysis of it, and the analysis must commend itself to us as an analysis of what we were already meaning by "truth". Now it may be said that our definition fails in this. It may be said 10 that we might equally have taken any other complex as the one which must exist if our belief is to be true. In fact, the association of belief and complex may be said to be too external in our definition. I do not think this is the case. Where non-permutative complexes are concerned, the complex formed of the objects of our belief seems as intimately associated with our belief as anything purely objective can be; and it seems quite evident that the truth or falsehood of a belief depends on something purely objective. Where permutative complexes are concerned, our process of obtaining associated non-permutative complexes was rather elaborate, and no doubt open to objection. One special objection is that, in order to regard the associated 20 complex as non-permutative, we have to regard its atomic constituents, XIC(y, X ZC 21', etc., as really its constituents, and what is more, we have to regard the corresponding propositions as constituents of the proposition "there is a complex I' in which x IC II', XzCzl', etc." This seems to demand a mode of analyzing molecular propositions which requires the admission that they may contain false atomic propositions as constituents, and therefore to demand the admission of false propositions in an objective sense. This is a real difficulty, but as it belongs to the theory of molecular propositions we will not consider it further at present. (2). It may be asked how truth and error can be distinguished if our 30 account is correct. We might retort by asking how they can be distinguished if any other account hitherto given is correct. Pragmatism, which prides itself on its supposed capacity for dealing with this question, is really, as we have seen in this chapter, in a much worse position than we are. Monistic idealism maintains a view from which it follows that we cannot know the truth of anything until we know the truth of everything. The inductive philosophy has never been able to explain how it ascertained the validity of inductive inference. And so on. But a tu quoque is not a very desirable form of argument, and I think a better reply is possible. To begin with, we have contended that it is possible to have acquaintance 40 with a complex. Seeing A and B together, we may have acquaintance with the complex "the similarity of A and B". This will, on our theory, assure us of the truth of the belief that A and B are similar. Thus our form of correspondence-theory, together wi th our doctrine of the direct perception
1
PART II, CHAP. V TRUTH AND FALSEHOOD
eling 'uth" f, we s not !fend :ions. alysis aning esaid must nplex : case. led of lief as uthor e perciated len to ciated uents, ave to )sition :1anda lission theresense. lecular
if our uished prides ,aswe .onistic ow the iuctive idityof Ie form
intance ce with ;sure us 'orm of ception
155
of complexes, seems unusually fitted to account for the possibility of belief which is knowledge, and for the way in which perception gives rise to such belief. It must be admitted, however, that no belief can amount to knowledge unless some beliefs are self-evident, in the sense that they are certainly true and are not inferred from other beliefs. The belief that A and B are similar, when we are directly acquainted with the similarity of A and B, will be a case in point. The conception of self-evidence is fundamental in distinguishing knowledge from belief and from true belief (which is not necessarily knowledge), Although we have dealt with belief and with truth and falsehood, we 10 have not yet dealt with knowledge, in the sense of knowledge of truths, as opposed to knowledge of objects, which is acquaintance. In a "theory of knowledge", it may seem strange to have postponed the consideration of knowledge so long. But in fact "knowledge" is a difficult and complicated concept, and many preliminaries are required before it can be successfully studied. In the next chapter, we shall begin this study by the consideration of self-evidence. (3). The last of the above objections to our theory, namely that nonmental "propositions" are, after all, indispensable, belongs to logic, not to theory of knowledge. It is, to my mind, much the most serious of the three 20 objections, and much the hardest to meet. I do not profess to be able to answer all the arguments in favour of "propositions" in this sense. I can only say that, to me personally, no such entities are visible, and the admission of such entities-which must be capable of falsehood as well as truth-runs counter to the rejection of unrealities, fictions, and mere possibilities which seems to me, on general grounds, necessary and vital to all sound philosophy. Until, then, the arguments in favour of non-mental "propositions" are presented in some more unanswerable form than any now known to me, I shall continue to reject them, and to believe that the apparent reasons in their favour are fallacious, even if I cannot always detect the 30 fallacy.
Chapter VI Self-Evidence
self-evidence is one of the most perplexing in the whole of Epistemology. Logic, psychology, and metaphysics all have something to contribute to it, but their various contributions are prima facie conflicting. It is easier to see that there must be such a thing as self-evidence than it is to see what it is. The endeavour to define selfevidence brings to a head the conflict between the objectivity of truth and the subjectivity of belief by which the scepticism of every age has been nourished. I do not pretend to be able to decide this secular conflict. But we shall find it not without interest and profit to explore the strength and weakness of the opposing forces. The broad definition of self-evidence is that it is knowledge which we possess independently of inference. So defined, it might incl ude know ledge of single objects by acquaintance; but although such knowledge is important in connection with self-evidence, it does not raise the problems which we have to discuss, and is therefore better excluded from the definition. We shall add, therefore, that self-evident knowledge is to be knowledge of propositions, not of single objects. In fact, since "acquaintance" serves for cognition of single objects, it will be convenient to define "knowledge" as cognition of propositions only, and is therefore only occurring in propositional thinking. The above definition would be admirable if we knew what is meant by "knowledge". But unfortunately the definition of "knowledge" is very difficult, and it seems highly probable that it must involve self-evidence: we may find that what we "know" is what is self-evident to us and what we infer from what is self-evident to us. "True belief" is not the definition of knowledge. We may have some wholly unwarrantable belief, such as that a particular horse will win the Derby, and our belief may turn out to have been true; but it is not on that account to be called knowledge. Some people believe that Home Rule will do good, some believe that it will not do good; one party in this opposition must be believing truly, but it would be absurd to say that either party had knowledge. We require, in knowledge, some power of logical resistance, some more than subjective stability against doubt, which may easily be absent in beliefs that merely happen to be true. Now this resistance and stability would seem always to rest, in the last
T
10
20
30
HE S lIBJECT OF
156
PART II, CHAP. VI SELF-EVIDENCE
157
resort, upon self-evidence. Hence we must define self-evidence without mentioning "knowledge" in our definition. Self-evidence cannot be defined except by reference to a particular subject, or to some psychological conditions. If anything is self-evident, it must be the knowledge immediately derived from sense; yet this is not selfevident until the moment of sensation. Hence it is not the nature of the proposition concerned that determines whether it is to be self-evident, for the proposition is just the same before sensation as it is after. It may be that there are discoverable kinds of propositions which are capable of selfevidence, i.e. that all self-evident propositions belong to certain kinds and 10 not to others; but actual self-evidence is a property which is relative to a given subject at a given moment, and does not belong to the proposition per se. It is better, therefore, to speak of self-evident beliefs than to speak of self-evident propositions. A self-evident belief must be independent of inference. I do not mean that it may not be ascertained by means of inference in the first instance, but that, when ascertained, it must be able to stand by itself without the help of the inference. Many abstract beliefs are obtained by inference from particular cases, and then become luminously obvious in their own right. This luminous obviousness is a sine qua non of self-evidence. 20 But besides this psychological characteristic, self-evident beliefs must also be true. We want self-evident beliefs to be the foundations of knowledge, and although some true belief is not knowledge, all knowledge is true belief. If we demanded merely the psychological characteristic of obviousness, we should escape all our difficulties, but we should not have found anything important to the theory of knowledge. We want, if possible, to find a class of beliefs which can be known to be true. For this purpose, we must so frame our definition of self-evidence as to insure that beliefs which have this characteristic are true. We do not want, however, to include truth explicitly in our definition of 30 self-evidence. If we do this, we are still left with the problem of discovering which of our beliefs possess truth and which do not, which is the very problem we hoped to solve by means of self-evidence. What we are seeking, therefore, is some characteristic, other than truth, which shall insure the truth of our beliefs without requiring that we should already know them to be true. Now luminous obviousness, by itself, seems, as a matter of empirical fact, to be insufficient to insure truth; at least, if it is to be sufficient, it must be very carefully defined and limited. Let us now turn to the logical side of our problem. If there is to be knowledge, there must be knowledge which is independent of inference. 40 This is the logical ground for saying that there can be no knowledge unless there is self-evidence. Let us see whether there is any escape from this necessity.
158
THEORY OF
K~O\"'LEDGE
The vulgar imagine that, in a science, every term ought to be defined and every proposition ought to be proved. But since human capacity is finite, what is known of a science cannot contain more than a finite number of definitions and propositions. It follows that every series of definitions and propositions must have a beginning, and therefore there must be undefined terms and unproved propositions. The undefined terms are understood by means of acquaintance. The unproved propositions must be known by means of self-evidence. It is sometimes sought to evade this conclusion by denying the linear 10 character of inference. Some-notably the monistic idealists-argue that all perfected inference is circular. Others, who advocate an inductive empiricism, contend that the apparent order of inference is never the real order, but that the truth of the premisses is only rendered probable by the conclusions to which they lead. It is not necessary, for our present purposes, to deny either of these views, for neither, we shall find, enables us to dispense with self-evidence. Let us take first the view that all inference is circular. In that case, to take a simplified scheme, we have (say) four propositions p, q, r, s, of which p implies q, q implies r, r implies s, and s implies p. We may also suppose, if 20 we like, that every other possible implication holds between these propositions. But all this affords no ground for believing any of them: however many mutual implications there may be, they may all be false. What is proved is that all are true if anyone of them is true. Thus if we have some reason, though not a conclusive one, for believing each separately, their mutual implications make the probability that they are all true greater than the antecedent probability that any separate one was true. This is important in many problems, but it is not relevant in our present problem. We must not, in our present problem, assume that we have an indubitable knowledge of propositions concerning probability, and thence prove that some prop30 ositions have a good chance of being true. The question as to how we acquired our indubitable knowledge of probability will have to be asked, and then we shall be forced ultimately to admit self-evidence as our only defence. The theory that all inference is circular, therefore, though it may alter the region in which self-evidence is to be sought, will not enable us to dispense with self-evidence as the ultimate source of our knowledge. We have now to consider the inductive argument, that the premisses are rendered probable by the conclusions to which they lead, and are not really the grounds on which we believe the conclusions. This contention rests on a confusion between logical and epistemological premisses. Given a body of 40 propositions which are all known to be true, or all known to be probable, and of which some are deducible from others, the best order, from a purely logical point of view, will be that in which there are the fewest and simplest premisses. But in theory of knowledge, where we wish to consider how
PART II, CHAP. VI SELF-EVIDENCE
f
e
r
:1
t
e
e
I, Y Y o
Iy
st
w
159
these propositions are known, the order is likely to be quite different. Our premisses will have to be self-evident, and it is not generally the case that the simplest logical premisses are so evident as some of their consequences. It may happen that, as the inductive empiricist contends, the simplest logical premisses are only rendered probable, not certain, by the self-evidence of the propositions which would be their consequences in a pure logical order. But this can afford no argument against self-evidence as the source of knowledge, since, if the whole body of propositions in question is to be accepted, self-evidence must belong to the propositions which are epistemological premisses and which give inductive probability to the purely logical premisses. Thus here again self-evidence remains epistemologically fundamentaL It may be said that the whole attempt to get behind belief to something more solid, and worthy to be called knowledge, is futile; that we believe and disbelieve at random, because we live a life of undisciplined impulse; but that, if we are seeking a philosophic detachment, we ought to realize that it is only to be found in a complete abstinence from belief, as from other forms of slavery to illusion. A less moralizing and more satirical scepticism may content itself with questions. How, it will ask, since you admit that you are sometimes wrong, can you know that you are not always wrong? How can you test beliefs except by other beliefs? And even if all your beliefs were coherent in one logical whole, which they probably are not, what illumination assures you that the truth must be coherent and the coherent may be true? Why should not the world be incoherent? Why not regard belief as merely an ineradicable propensity of the human animal, like eating and sleeping? To these questions there would seem to be no answer. The position ofthe sceptic who questions without denying is impregnable. All our arguments must appeal to some supposed common ground, something which the other side will admit; if nothing is admitted, argument ceases and refutation is impossible. The extreme sceptical position remains, therefore, one which is philosophically tenable. But the sceptical philosophy is brief; it begins and ends in questioning. By its nature, it cannot argue, or seek to establish any result, even its own tenability. Therefore its philosophical interest is soon exhausted, and we turn to the other hypothesis, according to which we do know some propositions. In short, we abandon the fundamentally questioning attitude, which is sometimes represented as alone truly philosophical, for the analytic attitude, which, in the main, accepts facts at their face value, and does not seek for a justification of the whole in something outside the whole. This attitude may not have any theoretical superiority over the other, but it at least leads to a more complicated philosophy, and therefore claims a longer attention. As soon as we adopt the analytic attitude, we are met by the fact that some
10
20
30
40
160
to
20
30
THEORY OF KNOWLEDGE
beliefs are recognized as erroneous, and that, therefore, some means of discriminating among beliefs is supposed to exist. It would seem that there must be logical grades among beliefs, and that those which survive the discrimination by which some are rejected must achieve a higher status than they had before. In short, within the system of beliefs, there must be means of giving more weight to some than to others, since without such means it would have been impossible to reject certain beliefs rather than others. If it could be shown that self-evidence, in some sense which insures truth, belongs to certain beliefs but not to all, the discrimination involved in rejecting some as erroneous might be justified. The necessity of self-evident beliefs if there is to be any knowledge is well stated by Meinong: I
In the circumstance that a conviction by no means guarantees the actuality of the proposition [Objektiv) which it apprehends, there lies a sort of danger for the whole of our knowledge. How then do we know about the actuality of the propositions in question? Obviously never otherwise than by judgments. But how do these judgments help us, if they in turn may be false just as easily as, perhaps more easily than, true? This does not, it is true, exclude the possibility of judging truly. But if it depended only on chance, that once in a way, so to speak, the right judgment and the right proposition should come together, then in the end it could also be only a chance when a man holds a iudgment which is true to be true. For the impulse, given with every judgment, to hold it to be true, belongs to false judgments no less than to true ones. Such consequences, which would be equivalent to abandoning all confidence in our judging, and so to the sacrifice of all knowledge, can only be avoided, as far as I see, by two presuppositions, first, that there are judgments in whose nature it lies to be true, secondly that we are capable of recognizing the truth-nature in such judgments by means of judgments of the same nature. Such judgments Meinong calls "evident", and he points out that it is evident that an evident judgment cannot be false. His view seems to be that "evidence" is an intrinsic property of certain judgments, and that judgments which have this property can be seen to have it by inspection, in fact that it is evident that they have it. By assuming that evident judgments are always true, we thus acquire a basis for distinguishing knowledge from mere belief. He spends little time on the analysis of evidence, but passes on to the enumeration of certain kinds of evident judgment.
40
1 Vber die Erfahrungsgrundlagen unseres Wissens (Berlin, 1906), p. 32.
PART II, CHAP. VI SELF-EVIDENCE
161
It should be observed that we cannot prove that evident judgments are true by appealing to the fact that it is evident that they are true. Such a proof would, of course, be a vicious circle. Evidence must be in itself a guarantee of truth, and cannot be used to prove that it is a guarantee. It must somehow lie in the nature of self-evidence that judgments which have this characteristic are free from the liability to error. Whether this is so, and how we can know it, are difficult questions. But before attacking these questions, it will be well to consider whether self-evidence can be analyzed, or whether there is some discoverable characteristic which is always associated with 10 self-evidence. Self-evidence may be regarded as consisting in some relation to acquaintance, or as a cognitive relation other than belief but implying it, or as a predicate of beliefs. Let us consider these views successively. (I). It may be said that a judgment is self-evident when it is contemporaneous with acquaintance with the corresponding complex. For example, if we see a patch of red surrounding a patch of white, the judgment that the red surrounds the white will be self-evident. This view has the merit that it secures the truth of self-evident judgments while yet making their selfevidence discoverable by inspection. According to this view, all selfevidence consists in analysis of what is given: when (say) a complex aRb is 20 given, the judgment that a has the relation R to b is self-evident. That judgments made under the above circumstances are self-evident is, I think, true, and it seems possible to maintain that their self-evidence consists in their being so made. As Meinong points out, it must sometimes, if not always, be evident that a judgment is evident, and this condition, so far as I can see, is fulfilled if the above is taken as a definition. If b is a belief, it is suggested that "b is evident" means "b is simultaneous with perception of the corresponding complex". Hence the belief that b is evident will be evident if it is simultaneous with the perception of b 's simultaneity with the perception of the complex corresponding to b. This may very well be the 30 case fairly often, and therefore it may fairly often be evident that a judgment is evident. There are, however, difficulties in the way of taking the above as a definition. Take the question of correspondence. We say that a judgment is evident if we perceive the corresponding complex. We do not say that it is evident if we perceive its correspondence with a certain complex, but only if we perceive the complex which does in fact correspond, though we may not know that it does. But this seems to leave self-evidence altogether too extraneous a property of judgments. It might quite well happen that we did in fact perceive the corresponding complex, without the judgment feeling 40 different from other judgments, unless perception of the corresponding complex causes the judgment to have some intrinsic property. But if any such intrinsic property is so caused, it may be capable of other causes. And I
162
10
20
30
40
THEORY OF KNOWLEDGE
think it is undeniable, on grounds of inspection, that evident judgments do feel different from other judgments. Take, again, the analytic judgments which we considered in Chapter II of this Part, such as "a is part of the complex aR b". If we derive the evidence of such judgments from perception of the corresponding complex, we have to suppose that, if they are to be evident, we must perceive the complex "a-part-of-aRb". This view, though perhaps not definitely refutable, makes analysis very complicated, and introduces great difficulties into the theory of the correspondence which defines truth. Such evident judgments as knowledge of the laws of contradiction and excluded middle would, I think, be very hard to fit into our formula for self-evidence. But as they, even in particular cases, introduce molecular propositions, we will not consider them at present. Let us leave the question whether our suggested definition of self-evidence is adequate undecided until we have considered other possible definitions. (2). We may hold that self-evidence is constituted by a special proposi. tional relation, different from belief but implying it. Just as attention is a kind of intensified acquaintance, so, it may be suggested, self-evidence is a kind of intensified belief. It may be said that, beyond ordinary belief, there is a kind of absolute certainty, leaving no possible room for doubt. Of course the obvious objection to this view would be that whatever we happen to be believing at the moment appears to us to be certain at the moment, if we have a real belief and not merely a degree of doubt which is all but certainty. But it would be said in reply that most belief is a mere unreflective assent, and that there is another kind of belief, which is capable of surviving critical scrutiny and adverse argument, because it makes what is believed strictly indubitable. Let us examine this view, first in the interests of the sceptic, then in the interests of the believer in knowledge. The view is one which the sceptic will welcome warmly. "Assuming", he will say, "that there is some hyper-dogmatic kind of judgment, which resists all our efforts to bring it under the domain of rational doubt, all the phenomena are fully accounted for. It is intelligible, to begin with, why the delusion that we possess knowledge is so widely diffused. Let a man once become the victim of an indubitability, and he will become utterly and absolutely convinced that he possesses a piece of truth from which no mere argument shall divorce him. If he drinks too much and thinks he sees rats, if he eats too little and thinks he sees angels, if he eats and drinks normally and thinks he sees tables and chairs, he is equally obsessed by indubitabilities." But it is also intelligible why the sceptic, though still victimized by the indubitable in his unprofessional moments, is able, on a calm survey, to perceive the groundlessness of even the most indubitable beliefs. For, though he may be unable to doubt what is to him indubitable, when it is
-.rI
PART II, CHAP. VI SELF-EVIDENCE
163
present to his mind, he may nevertheless succeed in doubting the general proposition "that the indubitable must be true". "Indubitability", he will say, "is a merely subjective property, belonging, at different times, to the most opposite beliefs. At one time, itisindubitable thatthe world is bad, that ingratitude and selfishness are all but universal, and that we alone maintain a heroic endurance and a stoical oblivion of our own outraged rights. At another time, under the influence of more sleep or better digestion, the contradictory of these propositions becomes equally indubitable." But the sceptic regards his moments of undoubtingness as the moments when instinct seizes him: it is only when he returns to the doubt whether what is indubitable is always true that he feels again the companionship of reason. "Action", he will say, "is what man shares with the brutes, and indubitable beliefs belong in essence with action. But thought is man's distinctive prerogative; and the more inconclusive it becomes, the more it is permeated by difficult doubt, the more completely is it purified from the taint of action, and the nearer it approaches to the ideal of contemplation. Let us recognize undoubtingness, therefore, as, like rage and hate and lust, one of the unfortunate passions to which our animal ancestry exposes us, but from which it is the business of the philosopher to free himself to the utmost possible extent." So far the rhetorical sceptic, whose desire for literary effect has led him into a somewhat naive attitude towards Darwin and delirium tremens. Let us now speak again in our own proper person, as advocates of knowledge. The truth in what the sceptic urges is that no merely subjective characteristic, such as indubitability, whether resulting from a special kind of beliefrelation or in any other way, can possibly give a guarantee of truth. It will account abundantly for men's belief that they have truth; but the better it accounts for this, the more it strengthens the sceptic's case, unless at the same time it gives a reason for certainty, and not merely a psychological account of certainty. Now a reason for certainty must involve reference to the facts. But an ascertainable reason for certainty must involve reference to ascertainable facts. Now the only way of ascertaining facts (except by judgment, which is liable to error) is by acquaintance. Hence if selfevidence is to be defined in such a way as to give a reason justifying certainty as regards our beliefs, it must be defined by means of acquaintance with facts connected with the beliefs. From this argument, so far as I can see, there is no escape. Hence, though indubitability should belong to selfevident judgments, their definition must not be by indubitability, but rather by some such method as we considered under our first heading. We must, however, still examine our third suggestion, that selfevidence is a predicate of some beliefs. There is a good deal to be said in favour of this view. It seems as if we could recognize a self-evident belief by the way it feels, without reference to anything outside itself. Meinong,
[0
20
30
40
164
THEORY OF KNOWLEDGE
though he does not enlarge on the analysis of self-evidence, seems to take this view. But the argument by which we refuted our second suggestion remains equally conclusive against our third. Truth is not an intrinsic predicate of judgments, for obviously their truth depends upon facts, not upon the character of the judgment as a mental occurrence. Thus if there is any intrinsic predicate of judgments which implies that they are true, it remains to be inquired how we can know this. It would be a petitio principii to argue that we know it because it is itself a judgment having the predicate in question. Hence the fact that the predicate in question is a guarantee of 10 truth must be somehow inferred from something which we know otherwise than through its possession of this predicate. There must, therefore, be some other definition of self-evidence, by means of which it can be known (if it is true) that this predicate insures truth. Thus once more we are brought back to some form of correspondence with perceived fact as the only possible source of self-evidence. It may be thought, however, that a similar argument could be employed against a definition of self-evidence by correspondence. How do we know, it may be said, that a judgment must be true ifit corresponds with fact? Is this self-evident? And if so, do you mean merely that it is perceived to corre20 spond with fact? But such an answer exposes you to the very petitio principii that you have urged against other definitions of self-evidence. The reply to this is that correspondence with fact is the definition of "truth", not merely a criterion. Hence when a judgment corresponds with fact, it is true by definition, not as the result of an inference. When a judgment is perceived to correspond with fact, it is indubitable; thus indubitability accompanies self-evidence, but does not constitute it, and may exist where self-evidence is absent. But owing to the fact that a self-evident judgment is indubitable, we do in fact firmly believe what is self-evident. It is in this way, through the fact that perceived correspondence causes belief, 30 that beliefs which deserve to be called "knowledge" arise. This explains at once why there are self-evident beliefs, and why self-evidence, where applicable, is a criterion of truth. Our argument so far has shown us (I) that self-evident judgments must be defined by some characteristic which is discoverable by inspection of what is given when such judgments are made, (2) that they must be so defined as to insure their truth, (3) that it is impossible to insure their truth except by reference to acquaintance with fact. Although, therefore, subjective certainty will, at least usually, belong to them, yet neither certainty nor any other subjective characteristic suffices to define them, and their definition 40 must be by reference to perception of some complex which insures their truth. We are thus brought back to our first theory, or something like it, as giving the right definition of self-evidence. It remains to examine several forms of this theory, and to consider which of them is the best.
PART II, CHAP. VI SELF-EVIDENCE
165
The general requisites of our arguments as to self-evidence are satisfied by any definition which makes self-evidence depend upon perception of some complex which only exists when the judgment is true. Besides what we call the corresponding complex, there are many other complexes which only exist when the judgmt'nt is true. One such complex suggests itself in this connection, namely the complex which consists of the correspondence of the judgment with the fact. Since this correspondence is what constitutes truth, we may be said, in this case, to perceive the truth of the judgment. Shall we say that this is what makes a judgment self-evident? There is some danger of confusing thefact that a judgment is self-evident with the knowledge that it is; and when we are said to know its self-evidence, it may be meant that we perceive it, or that we judge (truly) that it is self-evident, or that its self-evidence is self-evident to us. We have thus four things to distinguish: (1) the fact of self-evidence, (2) the perception of it , (3) the judgment of it, (4) the self-evidence of the judgment of it. It is only the fact of self-evidence that directly concerns us, since, when that is defined, the others follow. The only importance of mentioning the others is in order to remind us that they are different from the fact. Shall we say that a judgment is self-evident when we perceive the corresponding complex, or only when we perceive its correspondence with this complex? And in the latter event, need we perceive its correspondence with an actual complex, or only with some complex? It will be observed that "truth" is defined as correspondence with some complex; thus perception of the correspondence of a judgment with some complex is what constitutes perception of the truth of the judgment. If, therefore, it is this perception that defines self-evidence, then a judgment is self-evident when we perceive its truth. This view is one which seems worthy to be adopted if possible. The view that a judgment is self-evident whenever we simultaneously perceiVe the corresponding complex, which is the view we considered first, suffers from the defect that it makes self-evidence consist in such a very external characteristic. If we judge and simultaneously perceive the corresponding complex, without being conscious of the correspondence, it is difficult to see how this can give to our judgment the characteristic of self-evidence. It may be said that this constitutes the self-evidence of the judgment as a fact, but not the knowledge (in any of our three relevant senses) of its self-evidence. But I do not think this can be maintained. Mere simultaneity is in any case too external: it must be simultaneity in one experience. But even simultaneity in one experience cannot suffice: the relevance of the complex to the judgment must also be given in experience, otherwise the two fall apart, and we cannot account for the fact that a self-evident judgment feels different from one which is not self-evident. Suppose we are in a theatre before the beginning of the play: we shall believe that the curtain will rise, but this belief is not self-evident. At a certain
10
20
30
40
166
10
THEORY OF KNOWLEDGE
moment, we see it rising, i.e. we perceive the corresponding complex; at this moment our belief may become self-evident, but I think it only does do so if we perceive the correspondence of the curtain rising with our belief. Thus it seems that perception of the correspondence itself is essential to self-evidence. As to our second question, namely whether we must perceive correspondence with an actual complex, or only the fact of there being a corresponding complex, it is difficult to find any way of deciding it; for when we perceive the first we perceive the second, and it is doubtful whether we ever perceive the second without perceiving the first. There seem therefore to be no grounds of choice. I choose, arbitrarily, in favour of the second. Thus a judgment is self-evident when, at the time of making it, the person who makes it perceives its correspondence with some complex. Since this is defined as its truth, we arrive at the following definition:
Self-evidence is a property of judgments, consisting in the fact that, in the same experience with themselves, they are accompanied by acquaintance with their truth.
20
This definition, so far as I can see, satisfies all the conditions that we have found ought to be satisfied by a definition. And since there are self-evident judgments according to this definition, we have thus a means of making a beginning of knowledge as opposed to mere belief; for the definition shows that self-evident judgments are by nature incapable of falsehood, and therefore deserve the name of knowledge.
Chapter VII Degrees of Certainty
plex; at does do r belief. ~ntial to
e correbeing a g it; for loubtful . There avourof lking it, omplex.
n:
fact ,nied
we have '-evident flaking a •n shows ,od, and
T IS A commonplace that all our knowledge is in some degree liable to error, and that we are fallible even in our most dogmatic moments. Nevertheless, we feel far more certainty with regard to some beliefs than with regard to others; and this difference does not disappear with reflection, though the beliefs of which we are most certain in the end may be different from those of which we were most certain at the beginning. The question I wish to discuss in this chapter is whether there is any logical justification for different degrees of certainty, i. e. whether there is any ground for believing 10 some things more firmly than others. The question is important, because conflicts sometimes occur among our beliefs, and it is then a question which of them we shall abandon. If none of them are quite certain, it would be useful to know whether we have any reason to prefer the more certain; and this question requires an investigation of the question whether degrees of certainty have, or can have, any logical basis . We may introduce the discussion by considering what sort of certainty is actually derived from self-evidence. Of course, when our definition of self-evidence is strictly applicable, our belief has the highest possible degree of certainty, for we cannot perceive the truth of a belief unless the belief is 20 true. But before we can obtain a relatively fixed and isolated belief, expressed in words and registered mentally for future use, many mental operations have to be performed, and in all these there is some risk of error. Let us take the case of a judgment of sense, i.e. of a judgment such as "A is to the right of B", where A and B are patches of colour in one field of vision. Here sensation gives (by the help of selective attention) the complex A-tothe-right-of-B, and reflection enables us to perceive the correspondence of this complex with our judgment. Hence, if we perceive this correspondence and make our judgment, then our judgment is self-evident, and is entitled to the highest certainty. But in order to state and register our 30 judgment, A and B must be remembered; we must be sure that it was precisely those patches, not others near them, that were concerned; very 500n, A and B themselves will fade from memory, and we shall have to substitute descriptions of them, which involve possibly erroneous beliefs. In this way, any judgment of sense, however self-evident at first, is liable to change its character in unnoticed ways, and thus loses its complete certainty
I
167
ip
I 168
THEORY OF KNOWLEDGE
very quickly. In regard to logical judgments and judgments concerning universals, there is, it is true, much less ground for uncertainty when they have once been self-evident. That something has some relation to something, or that red and blue have that in common which makes us call them both colours, are propositions not referring to particular parts of time, and therefore as capable of self-evidence at one moment as at another. It is therefore theoretically possible to preserve their self-evidence by continued attention. But even in such judgments as these, there are remote possibilities of error if iO we look merely to the words which express the judgments, for it is possible that we may forget the meanings of the words and, trusting to our memory that they once expressed a truth, interpret them now in a way which is false. Such possibilities have little logical importance, but they have some practical importance, and they increase the urgency of our present problem, which is: To find a definition of degrees of certainty, which will give a reason, in cases of conflict, for preferring a more certain judgment to a less certain one. Our problem must not be confused with that of probability. That a proposition has a certain degree of probability, is a new proposition, which 20 we may believe with any degree of certainty from the highest to the lowest. Our problem is concerned in seeking something analogous to self-evidence as being nearly certain is analogous to being quite certain. We are not considering certainty about the probability of a proposition, but uncertainty about the proposition itself. Meinong l considers in this connection the case of memory. He points out that the trustworthiness of memory in general cannot be deduced from any other knowledge, although, of course, if memory in general is admitted, a particular case of memory may be tested by means of other cases. It follows that memory must be, in part at least, immediate knowledge, in the sense of 30 being uninferred, and yet is found, by lack of mutual consistency of recollections, to be liable to not infrequent errors. If, then, we are to have anything resembling knowledge concerning the past, it would seem that there must be something analogous to self-evidence in the power of yielding beliefs otherwise than by inference, but differing by the fact that the beliefs are not quite certain. We find, in fact, that our memories, quite independently of argument, are more or less certain: we feel convinced of what happened a moment ago, but very doubtful indeed concerning what we can barely remember. And we feel that this is reasonable; but why it is reasonable, is a very difficult question. 40 Meinong, after explaining the impossibility of proving the general trustworthiness of memory without assuming it, proceeds as follows (p. 70): I
Loc. cil., p. 68ff.
ip I PART II, CHAP. VII DEGREES OF CERTAINTY
169
It occurs to no one to await a demonstration before trusting his memory; everyone knows that he is justified in such trust even without a proof, i.e. on the ground, not of mediate, but of immediate evidence. Only this evidence stands, to begin with, in quite astonishing opposition to the fact that memory, not so very seldom, is deceptive, and that therefore a prudent person does not rely with full confidence on his recollections, but never quite loses sight of the possibility that he may be mistaken. But just here we have to take account of what gives judgments of memory such a characteristic place in our knowledge, and also primarily interests us in the present connection. These judgments are not made with certainty, or, more exactly: a man who does not wish to overstep his right in this matter must not make them with absolute certainty, though this happens often enough in practice. These judgments are, rather, in their nature only presumptions [Vermutungen], although, under favourable circumstances, such strong ones as to be hardly, or even not at all, perceptibly removed from certainty. The evidence, therefore, which belongs to them according to the above, is not evidence for certainty, but evidence for presumption. For there is, as can be seen e.g. in the region of calculable probability, the opposition of justifiable and unjustifiable judging, not only when we judge certainly, but also when we judge uncertainly. And something may be rightly presumed, which nevertheless does not happen, and so is false, which can of course never happen with justifiable certainty. This presumptive evidence, which involves the probability of the Objective of the judgment just as conclusive evidence involves its truth, I hold to be a fundamental fact of epistemology. f f e t
g 's It
n l-
The above passage brings out very clearly the epistemological importance of the problem which Meinong is discussing, and I do not see how to escape his conclusion, though I think his allusion to probability is a mistake, and I think we ought to exempt from uncertainty that part of the near past to which "immediate" memory reaches. It will be remembered that in Part I, Chapter VI, when we were discussing the perception of time, we derived our knowledge of the past entirely from immediate memory, not because this is the only source of our beliefs about the past, but because all other memory is fallible, and must, therefore, be supposed, since dual cognitive relations do not admit error, to begin with judgments, or at any rate with some kind of propositional thought. We will define the "immediate" past as the past which can be reached by immediate memory, and the "remote" past as the past which is not immediate. It may be doubted whether there is any sharp line of demarcation between them, but in any case there are portions of time
10
20
30
40
170
10
20
30
40
THEORY OF Kl\:OWLEDGE
well within the one, and portions well within the other. We have now to consider our beliefs concerning the remote past-i.e. all the past before the last thirty seconds or so. All that is to be said about "evident presumptions" may be said by reference to this special case, i.e. to our memory of the remote past. The theory of "evident presumptions" might seem to be forced upon us by Meinong's arguments. There are, however, some very grave objections of a general kind to the theory. When we were discussing self-evidence, we found that no theory was satisfactory which defined it by a purely psychological characteristic, without reference to fact; for, if there was no reference to fact, there could be no reason why self-evident judgments should be true. The same applies, it seems to me, to "presumptions". Memory, we are to suppose, consists of a number of "evident presumptions", which cannot all be true because they conflict inter se. But what reason have we for supposing that any of them are true? They seem true, but we know that this seeming is deceptive. To say that we are to accept, in cases of conflict, those we feel most sure of rather than the others, is wholly arbitrary. What has our feeling sure to do with the facts? We do, however, decide somehow, by means which it is hard to think wholly irrational, that some of our memories are mistaken, some true, and others still doubtful. If these means of deciding are not wholly irrational, evidence for presumption, like evidence for certainty, must be defined by some relation to facts with which we are acquainted. It cannot therefore be an ultimate analysis, however excellent it may be as a first account of what happens. The case of memory is, I think, much more complicated than Meinong supposes. In cases of erroneous memory, what seems to happen is that something imagined is mistaken for something remembered. Now we decided that imagination differs from memory and sensation by the fact that it does not imply (though it also does not exclude) a time-relation of subject and object. If, then, something which has actually been experienced becomes again an object of acquaintance, but without any given time-relation to the subject, it is now an object of imagination, not of memory. Since imagination seems limited, as regards simples, to objects which have been given in sense, all imagination is compounded of acquaintances with objects which have been experienced and therefore might be remembered, but in fact are not remembered. The question thus arises: what exactly is the mental occurrence which distinguishes remote memory from imagination? Time-relations to the subject seem to be twofold: there is pastness in general, and there are definite temporal distances. Within the immediate past, different distances can, I think, be immediately perceived: an event which happened a second ago seems less remote than one which happened five seconds ago. Temporal order within the immediate past could, in fact, be inferred from these relations of distance, if it were not otherwise known.
PART II, CHAP. VII DEGREES OF CERTAINTY
171
But our power of thus immediately perceiving temporal distances is very limited; beyond the immediate past, events remembered are simply past, and their greater or less distance from the present is a matter of inference. Now when we are considering a remote memory, we have to distinguish those cases (if any) in which the pastness is given in acquaintance, so that there is a given complex leading to the self-evident judgment "this is past", from those other cases where an object is given in an acquaintance which involves no time-relation to the subject, and is then judged, without the help of any time-acquaintance, to be an object which occurred in the past. In this latter case, the object, so far as it is given in acquaintance, is an object of 10 imagination; it is only through the addition of a judgment that it becomes an object of remote memory. Besides this distinction as regards the manner of our knowledge of the time-relation, there is another as regards the manner of our knowledge of the object. The object believed to be in the remote past may be not given in acquaintance, but only known by description. When we pass from imagination to memory by judgment, this seems usually the case: we do not believe that what existed was identical with what we imagine, but only that it may be described in terms of our image, by means ofthe kind of resemblance which commonly exists between sense-data and the images that we regard as images "of" those sense-data. This resemblance raises 20 difficult questions which do not concern our present problem; I shall therefore say no more about it now. We have thus, apart from perceptions of time-distance, four kinds of memory to distinguish, according as the pastness is given or judged, and according as the object is given or described. We will distinguish the first pair as "perceptive" memory and "judgment" memory; the second pair as "acquaintance" memory and "descriptive" memory. Thus: "Perceptive acquaintance-memory" consists in perceiving a complex of the form "A in the past", where A is known by acquaintance. "Perceptive descriptive memory" consists in perceiving a complex of the 30 form "the existence of the object of such-and-such a kind in the past", where the kind is usually determined by resemblance to an image which we have now. "J udgment-acquaintance-memory" consists in judgments of the form "A is in the past", where A is known by acquaintance, and the judgment is not derived from perceptive acquaintance-memory with A. "Judgment-descriptive memory" consists in judgments of the form "the object of such-and-such a kind is in the past", where the judgment is not derived from a perceptive memory. I do not know whether all these four kinds of remote memory actually 40 occur; but if memory and its fallibility are to be investigated, it is necessary to bear in mind that all these four kinds are possibilities to be considered. In considering whether these four kinds all occur, let us proceed back-
172
10
20
30
40
THEORY OF KNOWLEDGE
wards. Judgment-descriptive memory occurs, very obviously, in such a case as the following: I find my imagination haunted by a scene which has the dimness and vagueness commonly characteristic of imagination, and which therefore, as it stands, will not be thought to be identical with any set of sense-data. But suddenly it occurs to me that this scene is not mere imagination, but "represents" something that actually happened to me. Thus I judge it to be memory. It is descriptive memory because the image only resembles what happened; it is judgment memory because-at least so we assume-the first reference of our image to a past event is by a judgment. Such memory is quite peculiarly fallacious. Indeed the converse phenomenon to the above is not uncommon, where we mistake an imagined scene for a descriptive memory, and suddenly become aware--or at least begin to believe-that the scene is merely imaginary, and does not represent anything that really happened. Very remote memories, such as those of childhood, generally belong to this class. Unless corroborated otherwise, prudent people place very little reliance on such memories. Judgment-acquaintance-memory is more difficult to find an instance of. Perhaps, however, it occurs as part of judgment-descriptive memory when this uses a present image in its description. There probably are respects in which an image is expected to be identical with the thing it "represents"formal relations, spatial order, etc. For instance, an image of two men will contain the same sort of duplication of parts that belongs to the sense-datum called "seeing two men". Thus in respect of these elements which are identical, a judgment-memory which is on the whole descriptive will be an acquaintance-memory. There is another case which probably involves elements of judgment-acquaintance-memory, and that is the case of the illusion that what is happening has happened before. But this illusion does not as a rule go the length of belief; and if it did, we should only think that something exactly similar had happened before, and should therefore only have a description of the supposed previous occurrence. Similarly as regards the identity between image and sense-datum "represented", this does not extend to particulars, and therefore does not involve acquaintance with any actual particular which previously existed. Although, therefore, I see no reason to deny that judgment-acquaintance-memory occurs, I also see no reason to assert that it does. Perceptive descriptive memory will occur if, in such a case as we considered in connection with judgment-descriptive memory, it is at some moment self-evident that our image "represents" a past event. It may occur without this being self-evident in the sense defined in the previous chapter, if we ever perceive the complex "this representing some past event", where "this" will presumably be an image. That is to say, perceptive descriptive memory will occur in any case where we immediately perceive the SOft of
PART II, CHAP. VII DEGREES OF CERTAINTY
173
connection of our present image with the past which is involved in calling it an image "of" something past. This is the sort of experience which we might describe by saying that our image gives us a "feeling of pastness". I am inclined to think that this is the usual case of remote memory. But as a rule, the actual perception of the time-complex is very brief, and is quickly succeeded by the more stable and more easily revived judgment of memory. Thus this case of memory is, in practice, not easy to distinguish from judgment-descriptive memory. It will be seen that perceptive memory is infallible, while judgment-memory is fallible. Thus the facility with which the two may be confused makes errors of memory hard to avoid. But if, as I believe, perceptive descriptive memory is the usual case, it will follow that memory is usually trustworthy, and that, in a case which is not known for certain to be a case of perceptive memory, there is yet a probability in favour of its correctness. Perceptive acquaintance-memory occurs in immediate memory, which, however, contains also, if we were right in what we said above, a perception of temporal distance from the subject. The perception of temporal distance does not greatly concern our present inquiry, and may be ignored. We have, then, to ask ourselves whether perceptive acquaintance-memory extends beyond the narrow range of the immediate past. I think not. I think that this kind of memory is coextensive with immediate memory, and is indeed the analysis of immediate memory. Objects further in the past seem to be no longer given in acquaintance, but only indirectly accessible through images known to have a reference to the past. A particular with which we are acquainted is either a sense-datum, or a datum of immediate memory, or an image. It may happen, however, in exceptional cases, that the image has all the characteristics of a sense-datum, and is in fact identical with that "of" which it is the image. It may also happen in exceptional cases that immediate memory extends much further into the past than it usually does. These are purely empirical questions, the answer to which in no way affects our analysis. There is, however, one point in the psychology of memory which helps to explain what our account might otherwise seem incapable of explaining, namely the apparent gradual transition from immediate to remote memory. Between acquaintance-memory and descriptive memory there is, as a matter of logical form, an absolute gulf, with no possibility of gradual transition. But in the image which an event leaves behind it, there is a gradual transition: its intensity and brilliance and general resemblance to a sensedatum decays as time goes on, very rapidly at first, then more slowly. If it is by a description in terms of this image that we know the past sense-datum, it is intelligible that we think our knowledge of the past decays gradually. What really happens is that the present image which "represents" the past event comes gradually to be less and less like the past event, and thus there is
IO
20
30
40
174
THEORY OF
K~OWLEDGE
something which seems like a quantitative decay of acquaintance with the past, but is really acquaintance with an image growing progressively less like the past but known throughout to be "representative" of the past. Let us now return, after this long excursion, to the question of degrees of certainty, as illustrated by the case of memory. A memory-judgment may be based upon perceptive memory, and is then certainly true, or it may originate in judgment-memory, in which case it may be false. (When I speak of its "originating", I mean, on this occasion; of course, if it is true, it must have been originally an acquaintance, but that does not enter into our 10 present question.) Let us dismiss the judgment-memory, as not worthy to be called "knowledge" even when it happens to be correct. (It is to be observed that, when I speak of memory, I include no beliefs about the past which are inferred from other beliefs. Inferred beliefs about the past may be knowledge, but in that case, among the premisses from which they are inferred, there will always be some memory-judgments based on perceptive memory.) Thus our store ofuninferred knowledge concerning the past will consist of memory-judgments based on perceptive memory. Such judgments, we know, are true. But does that give us any right to rely implicitly upon this or that memory-judgment? In other words, can we ever be sure 20 that a given memory-judgment is based on perceptive memory, not on judgment-memory? The question of degrees of certainty, which has escaped us so long, reappears here. For a moment, perceptive memory gives us assurance. If we can reach the memory-judgment while perceptive memory still lasts, we have certainty as regards the memory-judgment. But if the perceptive memory is very fugitive, as it often is, our memory-judgment will soon come to be based, not upon the original perceptive memory, but upon remembering the perceptive memory. This remembering may itself be of any of our four kinds. If it is perceptive acquaintance-memory, we may still feel 30 quite certain of what we originally remembered; but in any other case, we ought to feel a certain doubt. In the case of judgment-memory, this is obvious. In the case of perceptive descriptive memory, where we remember our memory by means of an image which gradually grows less like the memory which it "represents", the image gives less and less information about the original memory, and therefore must give less and less assurance as to the nature of the fact originally remembered. And apart from the complication introduced by remembering a memory, in all cases of descriptive memory based on images, our knowledge as to what is remembered has a certain vagueness, due to the vagueness of the relation of "representing" 40 which connects the image with the past fact. This seems to account for our indestructible certainty that there was a past, which we know to have existed in virtue of memory, in spite of the fact that particular memories often prove to be wrong. The representative nature of the image, according to our
PART II, CHAP. VII DEGREES OF CERTAINTY
e f :I
e 'r
e n
e e )-
lS
Ir
d
'e Ir
175
theory, is immediately known in descriptive perceptive memory; no doubt the accuracy of the representation must be known within limits, if memory by images is to give any information. But the limits may sometimes be wide, and thus any unduly precise statement as to what is remembered will be uncertain. It would seem, therefore, that the ultimate source of the uncertainty which is still present in perceptive memory by images, is the vagueness of the representative character of images. Degrees of certainty, therefore, in this case at least, are not an ultimate property of judgments, but are derived from vagueness. We ought not to pass an uncertain judgment that such-and-such a thing happened, but a certain judgment that something 10 like such-and-such a thing happened. From this it is theoretically possible to pass to probability-judgments. Thus it would seem that what we took to be an uncertain judgment about something is really a certain judgment about something else. The subject of vagueness is of some importance, and has not been sufficiently considered. Let us endeavour to define it, taking the case of memory-images for purposes of illustration. Suppose the question comes up how tall a certain absent person is. We may endeavour to decide the question by calling up an image of him. It may be easy from the image to decide that he is over 5 ft. 8 in. and under 6 ft. 2 in. 20 But it may be impossible to decide from the image whether he is over or under 6 ft., though we may feel "nearly certain" that he is under 6 ft. Let us assume that there are people present whose heights are respectively 5 ft. 8 in., 5 ft. 9 in., 5 ft. loin., 5 ft. II in., 6 ft., 6ft. I in. and 6 ft. 2 in. We will call these people A, B, C, D, E, F, G. Our image enables us to judge that the absent man is taller than A and shorter than G . We think he is taller than B and shorter than F, but do not feel sure; D seems the most likely height, but B, C, E and F are possible. How is this state of affairs to be analyzed? I t might be said that our image is of a certain definite height, but we are not sure that the man is of the same height. This, however, does not seem to fit 30 the facts. The image will be more nearly of some exact height in a good visualizer than in a bad one, but even in the best visualizer it will not be distinguishable with any certainty from either of two given heights both of which can be easily distinguished from a height half way between them. It has therefore a certain vagueness which does not belong to sense-data. What we call "height" in an image is perhaps not the same as height in a sense-datum, but it has some relation to height in a sense-datum, since it enables us to pass some judgments as to the height of the absent man. This relation, whatever it is, is part of what makes the image "representative". It would seem that there are as many heights of images as of sense-data, for it 40 would be a different image that would assure us that the absent man was between 5 ft. 9 in. and 6 ft. 3 in. But each image is capable of "representing" a number of different heights. Perhaps there is one which it represents best,
176
THEORY OF KNO\X"LEDGE
while those on either side are represented gradually less and less well. However this may be, the relation of "representing", which holds between images and sense-data, is not one-one; a whole stretch of objects may be represented by a given image, and a whole stretch of images may represent a given object. This fact seems to constitute the logical analysis of "vagueness". What has been said about height is obviously applicable to any similar series, such as colours, tones, etc. What has been said above about the uncertain judgments of memory will be found, I think, to apply equally to all other cases of uncertainty. That is 10 to say, there is no such thing, if we were right, as an uncertain judgment: there is only a judgment of uncertainty. In the case of the man whose height we try to decide from our image of him, it is uncertain whether he is over or under 6 ft. But this is not an uncertain judgment that he is over 6 ft., together with an uncertain judgment that he is under 6 ft. It is a single judgment, capable of just as great certainty as any other. If this conclusion is sound, ll'leinong's notion of "presumptions", as something falling short of the decidedness of judgments, is not ultimate, but resolves itself into what we may call the "judgment of approximation", or sometimes into the "judgment of probability", which are distinguished by the nature of their 20 objects, not by the manner of judging. I do not feel any great confidence in this conclusion, but on the whole it seems more probable than any other. It should be observed that the question of error is wholly distinct from our present question. Judgments in which we have the most absolute confidence sometimes turn out to have been erroneous. As we saw in discussing self-evidence, error cannot be excluded by any purely subjective characteristic such as the feeling of certainty, but only by that immediate contact with facts which we obtain in acquaintance. The question of criteria for the exclusion of error is not one with which we can deal at the present stage; but from what has been said already, it is obvious that such criteria must have 30 reference to acquaintance, and must be undiscoverable so long as we remain within the sphere of judgment. Before passing on to the consideration of molecular thought, let us sum up what has been said in this Part on the subject of atomic thought. An atomic complex, broadly, is one in which there is a single principal relating relation, i.e. the complex can be analyzed into certain terms (which may be complex) related by a single relation. An atomic propositional thought is a complex formed by a multiple relation of a subject to certain objects, where nothing is involved in the objects which may not occur in an atomic complex, but the "form" of some atomic complex does occur, while 40 no "form" of any molecular complex occurs. What is called "understanding a proposition" is a relation of a subject to certain objects, which are 1) the form of certain atomic complexes entities of the same logical kinds as the constituents of such complexes,
PART II, CHAP. VII DEGREES OF CERTAINTY
177
sufficient in number and kind to form one such complex. A given proposition will be the fact (if it is a fact) of there being a complex of the form of an understanding, where the objects are given, but the subject and the relating relation are arbitrary. The question of analysis, we found, is complicated by the doubt as to whether we can be acquainted with a complex without being acquainted with its constituents. What seemed empirically certain was, that we may be acquainted with a complex without being able to discover that we are acquainted with its constituents; but it is not possible to assert positively that we are ever not acquainted with its constituents. We found that there are two ways of perceiving a complex, which we called respectively simple and complex perception. In simple perception, we attend to the complex itself, and do not attend to its parts, whether or not we are acquainted with them. In complex perception, we attend to the parts, and are acquainted with the whole without attending to it. It is by means of complex perception that we are able to give a complex name to a complex, composed of the names of its constituents suitably combined. When we have become able to give a complex name, the judgments that this and that and the other are constituents of the complex are self-evident. In considering the understanding of propositions, a specially interesting case is afforded by the atomic propositions of pure logic, which have no constituents; in this case, understanding is a dual relation, the object-term being a pure form. This fact seems to be connected with the self-evidence of logical propositions. The understanding of such propositions is the logically simplest kind of understanding. Next to this comes the understanding of propositions which merely assert that there are instances of such-and-such a universal. The understanding of unsymmetrical atomic complexes, such as "a-before-b", raises special problems, and appears to be not an atomic understanding. In connection with belief, which we considered next, we emphasized the necessity of regarding it, and propositional thought generally, as consisting in a multiple relation, not a dual relation, and as thus radically different from conception or any form of acquaintance. The assimilation of belief to conception, we found, has been the source of many false theories, from Hume onwards. It has been encouraged by a wrong analysis of "existence", and has produced inextricable confusions in the theory of error. Belief, we decided, and disbelief also, is a relation producing precisely the same form of complex as is produced by understanding, and raising no new logical problems. Truth and falsehood, which we considered next, we defined by the presence, in the case of true belief, of a certain definable correspondence with a complex consisting of the objects of the belief, while in the case of falsehood there is no such complex. The chief difficulty lay in defining the
10
20
30
40
r
178
10
20
THEORY OF KNOWLEDGE
correspondence in the case of unsymmetrical homogeneous complexes, but we overcame this by constructing associated non-homogeneous complexes which exist when the original complexes do but not otherwise. We were thus able to state our theory finally in the simple form: A belief is true when its object-terms form a complex; if not, it is false. This theory we endeavoured to strengthen by a criticism of various other theories which have been proposed. The subject of self-evidence, which we considered next, is important as introducing us for the first time to something which may be called knowledge as opposed to mere belief. We found that there must be some beliefs of whose truth there can be no doubt, though they are not obtained by inference; if this were not the case, there could be no such thing as knowledge. Such beliefs, we found, must be recognizable by inspection, and yet not defined by any purely psychological characteristic. They must therefore involve reference to facts with which we are acquainted. The definition we arrived at is: A belief is self-evident when we are acquainted with its truth, i.e. with the fact that there is a complex composed of its object-terms. We saw that there are beliefs which are self-evident in this sense, and that sometimes it is self-evident that they are self-evident. Finally we considered degrees of certainty, and decided, though with some hesitation, that this is not an ultimate notion, but that what appears to be an uncertain judgment is really a certain judgment differing as regards its objects. This conclusion we reached largely by an analysis of memory which is not immediate, and of the way in which a memory-image represents a past datum.
{es, but nplexes 7e were le when we ench have
,nam as owledge eliefs of ined by ; knowland yet ~erefore
irion we ts truth, ·ms. We md that
,gh with 'pears to
gards its rywhich ltS a past
Bibliographical Index
THE WORKS REFERRED to by Russell, and by the editors using the author-date system, are cited here in full. The year in italics following the author's name is that of the edition described, or else that of first publication. Full bibliographical data are not provided for pre-I 800 titles. After each citation is an index of the pages on which the work is mentioned. Page numbers in roman type refer to Russell's text; page numbers in italics to editorial matter. The bulk of Russell's working library is housed in the Russell Archives at ,\kMaster 1:niversity. The phrase "(Russell's library.)" indicates relevant volumes in that library. The Bibliographical Index does not list unpublished correspondence referred to in the Introduction.
AYER, A.J., W.C. KNEALE, G.A. PAUL, D.F. PEARS, P.F. STRAWSON, G.J. WARNOCK AND R.A. WOLLHEIM, 1956. The Rez'olUlion in Philosophy. London: Macmillan; New York: St ,'.lartin's Press. Referred to: xiiin. BLACKWELL, KENNETH, 1973. "Our Knowledge of Our Knowledge". Russell, no. 12: 11-13. Referred to: ixn. 1974. "Wittgenstein's Impact on Russell's Theory of Belief'. Unpublished M.A. thesis, McMaster University. Referred to: viiin. J98J. "The Early Wittgenstein and the Middle Russell". In Block 1981. Referred to: ixn. AND EUh\IWIH RAAISDE:-.J EA.\1Es, 1975. "Russell's Unpublished Book on Theory of Knowledge". Russell, no. 19: 3-14,18. Referred to: viiin. BLOCK, IRVING, ed., 1981. Perspeaives on the Philosoph)) of Vll1lgenslein. Oxi()rd: Basil Blackwell. Referred to: ixn. Bf(o/w, C.D., 1924. "Critical and Speculative Philosophy". In COnlemporary Bntish Philosophy: Personal Slatements (FIrs! Series). Edited by J .1I. 179
BIBLIOGRAPHICAL INDEX
180
Muirhead. London: George Allen & Unwin; :-Jew York: Macmillan. Referred to: ·Vli. CHISIIOL.\1, RODl,RICK "'1., 1974. "On the Nature of Acquaintance: A Discussion of Russell's Theory of Knowledge". In Nakhnikian 1974. Referred to: viiin. Cl.ARK, RONAl.D W., 1975. The LIfe 0/ Bertrand Russell. London: Jonathan Cape and Weidenfeld & Nicolson. Referred to: viiin. COSTELLO, HARRY T., 1957. "Logic in 1914 and Now". The Journal o/Philosophy, 54: 245-64. Referred to: xxiv. COSTEl.LOE, KARI"I, 1914. "An Answer to Mf. Bertrand Russell's Article on the Philosophy of Bergson". The AIonist, 24: 145-55. (Russell's library.) Referred to: xxiii. DEWEY, JOHN, 1916. Essays in Experimental Logi./.'. Chicago: The University of Chicago Press. (Russell's library.) Referred to: xxvin. EAMES, ELIZABETH RA.\1SDE"I, 1969. Bertrand Russell's Theory 0/ Knowledge. London: George Allen and Unwin. Referred to: viiin. 1975. "Phillip E.B. Jourdain and the Open Court Papers". [CarbS, 2:
IOH2. Referred to: ixn. 1979. "Response to Mf. Perkins". Russell, nos. 35-6: 41-2. Referred to: ixn. See also BI.ACKWEl.l., KENNETH. AKD ELlZAllETH RAMSDEN EAMES. ELIOT, T.S., 1914. Unpublished seminar notebook. Houghton Library, Harvard. Referred to: xn., xxiiin.-xxivn. Fl.ORENCE, P. SARGANT, 1977. "Cambridge /909-1919 and Its Aftermath". In G.K. Ogden: A Collectit'e Memvir. Edited by P. Sargant Florence and J .R.L. Anderson. London: Elek Pemberton. Referred to: xxiiin. GRIFFIN, NICHOl.AS, 1980. "Russell on the Nature of Logic (1903-1913)" . .'l)mthese, 45: 177-88. Referred to: ixn. 1982. "Russell's Multiple-Relation Theory of Judgment". Unpublished paper read at Foundations of Logic Conference, University of Waterloo. Referred to: ixn. HOl.T, EnwlK B., \XrALTER T. MARVIN, WILLIAM PEPPERELL MONTAGUE, RALPH BARTON PERRY, WAl.TER B. PITKIN ,\ND EDWARD GLEASON SPALDlKG, 1912. The l,'ew Realism: Cooperative Studies in Philosophy. New York: Macmillan. Referred to: 16n., 24. HeME, DAVID, 1888. A Treatise 0/ Human Nature. 2 vols. Edited by T.H. Green
BIBLIOGRAPHICAL INDEX
181
and T.H. Grose. London: Longmans, Green. (Russell's library.) Referred to: 136,137,138,139. JAMES, WILLIAM, 1890. The Principles of Psychology. 2 vols. London: Macmillan. (Russell's library: 1891 [i.e. 3rd] impression.) Reprinted in his 1975~. Referred to: 17,55, 72n., 140n. 1912. Essays in Radical Empiricism, London, New York and Bombay: Longmans, Green, (Russell's library.) Reprinted in his 1975-. Referred to: ISn" 17, 18, 19,20,27. JOACHIM, HAROLD H., 1906, The Nature of Truth. Oxford: at the Clarendon Press, Referred to: 149, KEYNES, JOHN MAYNARD, 1921, A Treatise on Probability. London: Macmillan. (Russell's library.) Referred to: xxixn. LACKEY, DOlJ(;LAS, 1981. "Russell's 1913 Map of the Mind". Midwest Studies in Philosopl!.v, 6: 125-42. Referred to: ixn. LENZE:-.I, VICTOR, 1914. "Epistemology", Unpublished seminar notebook. RA REC. ACQ. 133b. Referred to: xn., xxii, xxiii, xxiv, xxv, xxix, xxx, xxxiii, xxxvi. 1914a, "A Theory of Judgment", Unpublished term paper for B, Russell's epistemology seminar, Harvard University. Bancroft Library, University of California, Berkeley; copy in RA (REe ACQ. 133a). Referred to: xn" xxv. 1971. "Bertrand Russell at Harvard, 1914". Russell, no, 3: 4-6. Referred to: xxiz'n, MACH, ERNST, "A Theory of Heat". Unpublished translation, in Open Court papers, of Die Principien der Wdrmelehre; historisch-kritisch entwickell. Leipzig: Johann Ambrosius Barth, 1896. Referred to: xxviin, 1897, ContribUlions to the Analysis of the Sensalions. Translated by C.M. Williams. Chicago and London: Open Coun. 1st German ed., 1866. Referred to: ISn., 16. MAXWELL, GROVER, 1974. "The Later Bertrand Russell: Philosophical Revolutionary". In Nakhnikian 1974. Referred to: viiin. M.cGUINNESS, BRlAN, 1972. "Bertrand Russell and Ludwig Wingenstein's 'Notes on Logic' ". Revue internationale de philosoph ie, 26: 444-60. Referred to: viiin., xxin. 1974. "The Grundgedanke of the Tractatus". In Understanding Wiugenstein. Edited by Godfrey Vesey. (Royal Institute of Philosophy Lectures, Vol. i.) London: M.acmillan. Referred to: ixn. MEINONG, ALEXIUS, 1899. "Uber Gegenstlinde h6herer Ordnung und deren
182
BIBLIOGRAPHICAL INDEX
Vcrhaltniss zur inncrcn Wahrnehmung". Zeilschriji Jur Psychologie und Physiologic des Sinnesorgane, 21: 182-272. Reprinted in his 1968-78, Vol. 2. Referred to: 41,42. 1906. Ober die ErJahrungsgnmdlagen unseres W'issens. i,Abhandlungen zur Didaktik und Philosophie der Naturwissenschaft, Band 1, Hett 6; Sonderhcfte def ZeitschriJt Jur den physikalischen und chemischen Unterricht.) Berlin: Julius Springer. Reprinted in his 1968-78, Vol. 5. Referred to: 160, 168n., 169. MON'r.~(;UE, WILLL\M PEPPERELL, 1912. "A Realistic Theory of Truth and Error". In Holt el al., 1912. Referred to: 24. NAKHSIKIAN, GEORGE, ed. 1974. Bemand Russell's Philosophy. London: Duckworth. Referred to: viiin. PEARS, D.F., 1975 "Russell's Theories of Memory". In his 1975b. Referred to: ixn. 1975a. "Wittgenstein's Treatment of Solipsism in the Tractalus". In his 1975b. Referred to: ixn. 1975b. Questions in the Philosophy oJAfind. London: Duckworth. Referred to: ixn. 1977. "The Relation between Wingenstein's Picture Theory of Propositions and Russell's Theories of Judgement". Philosophical Re~jie'lo, 86: 177-96. Reprinted in W'iltgenstein: Sources and Perspectives. Edited by e.G. Luckhardt. Ithaca, N.Y.: Cornell University Press, 1979. Referred to: ixn. 1979. 'Wingenstein's Picture Theory and Russell's Theory of Knlrwledge". In Willgenslein, the Vienna Circle and Critical Rationalism: Proceedings of ehe Third Internalional Wiugenstein Symposium. Edited by H. Berghel, A. HUbner and E. Kohler. (Schriftenreihe der Wittgenstein-Gesellschaft.) Vienna: Holder-Pichler-Tempsky. Referred to: ixn. 1981. "The Logical Independence of Elementary Propositions". In Block 1981. Referred to: ixn. 1981a. "The Function of Acquaintance in Russell's Philosophy". Symhese, 46: 149--66.
Referred to: ixn. JR., R.K., 1979. "Russell's Unpublished Book on Theory of Knowledge". Russell, nos. 35-6: 37-40. Referred to: ixn. PERRY, RALPH R,\RlO~, 1912. Presem Philosophical Tendencies: A Crieical Survey oJ N arura/ism, Idealism, Pragmatism and Realism together Wilh a Synopsis oJ ehe PERKI~S,
BIBLIOGRAPHICAL IN'DEX
183
P/ulosophy of William James. London, Bomba\" and Calcutta: Longmans, Green. Referred to: 16n., 21, 29, 30. RnsELL, BI:.RTRA~D ARTHL'R \"{'ILU,',,\1. .;RD EARL, 1903. The Principles of Mathemalics. Cambridge: at the University Press. 2nd ed., London: George Allen & Unwin, 1937. Referred to: 84n., 87n. 1904. "Meinong's Theory of Complexes and Assumptions". ,Hind, n.s. 13: 204-19, 336-54, 509~24. Reprinted in his 1973. Referred to: 153. 1905. "On Denoting". Mind, 14: 479-93. Reprinted in his 1956. Referred to: xii. 1905a. Review of Meinong 1904. Mind, n.s. 14: 530-8. Reprinted in his 1973. 1906. "What is Truth?" Independent Review, 9: 349-53. Review of Joachim 1906. 1906a. "The Namre of Truth". },Jind, n.s. 15: 528-33. Review of Joachim 1906. 1907. Review of Meinong 1907. Mind, n.s. 16: 436--9. Reprinted in 1973. 1907a "On the Nature of Truth". Proceedings of ehe Aristotelian Sociely, n.s. 7:28-49. Sections I and II reprinted as "The M.onistic Theory of Truth" in his 1910. 1908. "Transatlantic 'Truth' ". Albany Review, n.s. 2: 393-410. Reprinted as "William James's Conception of Truth" in his 1910. 1910. Philosophical Essays. London, New York and Bombay: Longmans, Green. New ed., London: George Allen & Unwin, 1966. 191Oa. "On the Nature of Truth and Falsehood". In his 19/0. 191Ob. "La Theorie des types logiques". Rwue de melaphysique el de morale, 18: 263-301. Reprinted in English in the Introduction to Vol. I of Whitehead and Russell 1910-13 and in Russell 1973. 19/0-13. See Whitehead and Russell 1910-13. 1911. "L'Importance philosophique de la logistique". Rez'ue de melaphysique el de morale, 19: 281-91. See his 1913 for a translation. Referred to: xin., xliii. 1911a. "Knowledge by Acquaintance and Knowledge by Description". Proceedings ofehe An'slolelian Socie~v, n.s. 11: 108-28. Reprinted in his 1918. Referred to: xliii, 36n. 1911b. "Le Realisme analytique". Bulleein de la Societe franfaise de philosophie, 11: 53-82. Referred to: xn., xliii. 1911c. "Sur les Axiomes de l'infini et du transfini". Sociile mathemalique de France: Comples rendus des seances, no. 2: 22-35. Translated into English as "On the Axioms of the Infinite and of the Transfinite" in 1. Grattan-
184
BIBLIOGRAPHICAL INDEX
Guinness. Dear Russell-Dear Jourdain: A Commentary on Russell's Logic, Based on His Correspondence with Philip Jourdain. London: Duckworth, 1977. Referred to: xliii. 1912. The Problems of Philosophy. (Home University Library of Modern Knowledge, No. 35.) London: Williams and Norgate, [n.d.]. Referred to: vii, xi, xii, xiii, xix, xliii. 1912a. "On the Relations of Universals and Particulars". Proceedings of the Aristotelian Society, n.s. 12: 1-24. Reprinted in his 1956. Referred to: xliii. 1912b. "The Philosophy of Bergson". The Monist, 22: 321-47. Reprinted as "Bergson", Bk. 3, Chap. 28, in his A History of Western Philosophy. New York: Simon and Schuster, 1945. Referred to: xliii. 1912c. "The Essence of Religion". Hibbert Journal, II: 46--62. Reprinted in The Basic Writings of Bertrand Russell, 1903-59. Edited by Robert E. Egner and Lester E. Denonn. London: George Allen & Unwin, 1961. Referred to: xi, xliii. 1912d. Review of James 1912. Mind, n.s. 21: 571-5. 1912e. "On Matter". Paper read at Cardiff. Unpublished manuscript. RA 220.011360. Referred to: xliii. 1912f. "What is Logic?" Unpublished manuscript. RA 220.011430. Referred to: xliii. 1913. "The Philosophical Importance of Mathematical Logic". Translation of his 191I by Philip Jourdain. The Monist, 23: 481-93. Reprinted in his 1973. Referred to: xin. 1914. Our Knowledge of the External World as a Field for Scientific Method in Philosophy. Chicago and London: Open Court. Revised ed., London: George Allen & Unwin, 1926. Referred to: vii, x, xiii, xviii, xxi, xxiv, xxvi, xxvii, xxviii, xlvii. 1914a. "Mysticism and Logic". Hibbert Journal, 12: 780-803. Reprinted in his 1918. Referred to: xiiI, xxiii, xlvi. 1914b. "On the Nature of Acquaintance": [Part I.] "Preliminary Description of Experience". The Monist, 24 (Jan.): 1-16. Reprinted in his 1956 and as Part I, Chap. I, of this volume. [Part] "II. Neutral Monism". The Monist, 24 (April): 161-87. Reprinted in his 1956 and as Part I, Chap. II, of this volume. [Part] "III. Analysis of Experience". The Monist, 24 (July); 435-53. Reprinted in his 1956 and as Part I, Chap. III, of this volume. Referred to: xxi, xxii, xxiii, xxvi, xxvii, xxix, xxxi, xxxiii, xlvi.
BIBLIOGRAPHICAL INDEX
1914c. "Definitions and Methodological Principles in Theory of Knowledge". The Monist, 24 (Oct.): 582-93. Reprinted as Part I, Chap. IV, of this volume. Referred to: xviiin., xxiv, xxvin., xxix, xxxi, xxxiii, xxxiv, xlvii. 1914d. "The Relation of Sense-Data to Physics". Scientia, 16: 1-27. Reprinted in his 1918. Referred to: xxiii, xxv, xxviii, xxix, xlvi. 1914e. "Preface". In Henri Poincare. Science and Method. Translated by Francis Maitland. London, Edinburgh and ~ew York: Thomas Nelson and Sons, [n.d.]. Referred to: xxiii. 1915. "Sensation and Imagination". The Monist, 25 (Jan.): 28-44. Reprinted as Part I, Chap. V, of this volume. Referred to: xxvi, xxviii, xxix, xxxi, xlvii. 1915a. "On the Experience of Time". The ,'.ioniSE, 25 (April): 212-33. Reprinted as Part I, Chap. VI, of this volume. Referred to: ix, xviiin., xxiv, xxvi, xxviii, xxix, xxxiii, xlvii. 1916. Justice in War-Time. Chicago and London: Open Court, 2nd ed., 1917. Referred to: xxvii. 1916a. Principles of Social Reconstruction. London: George Allen & Unwin . u.s. ed. as Why Men Fight: A Method of Abolishing the International Duel. New York: The Century Co., 1917. Referred to: xxt'iin. 1918. Mysticism and Logic and Other Essays. London, New York and Bombay: Longmans, Green. 1918-19. "The Philosophy of Logical Atomism". 8 lectures. The Monist, 28 (Oct. 1918): 495-527; ibid., 29 (Jan.-July 1919): 33-63, 19~222, 345-80. Reprinted in his 1956. Referred to: 'oii, viiin., xxxv, xlvii. 1919. Introduction to Mathematical Philosophy. London: George Allen & Unwin; New York: Macmillan. Referred to: xxxvin. 1919a. "On Propositions: What They Are and How They Mean". Aristotelian Society Supplementary Volume, 2: 1-43. Reprinted in his 1956. Referred to: xxxvi. 1919b. "Professor Dewey's 'Essays in Experimental Logic'''. The Journal of Philosophy, Psychology and Scientific Methods, 16: 5-26. Review of Dewey 19]6. Reprinted in Dewey and His Critics: Essays from the Journal of Philosophy. Edited by Sidney Morganbesser. New York: The Journal of Philosophy, 1977. Referred to: xxvin. 1921. The Analysis of Mind. London: George Allen & Unwin. Referred to: xxxvi .
.ogic, orth,
Idem
if the ed as New
ed in :gner
.• RA
Ilion his
1
od in 'orge
~d
in
-16. 1
his
~d
185
in
l· ,·~i
~ .
'
.
186
BIBLIOGRAPHICAL INDEX
1922. "Introduction" to Tractatus Logico-Philosophicus. German translation in Wittgenstein 1921; in English, with revisions, in Wittgenstein 1922. Referred to: xxxvi, xxxvim. 1927. The Analysis of Matter. London: Kegan Paul, Trench, Trubner; New York: Harcourt, Brace. Referred to: viiin. 1944. "My A·1ental Development". In Schilpp 1944. Referred to: xivn. 1956. Logic and Knowledge: Essays, /901-1950. Edited by R.C. Marsh. London: George Allen & Unwin. Referred to: x, xxxvn. 1959. My Philosophical Development. London: George Allen & Unwin. Referred to: ,)iiin., ixn. 1967. The Autobiography of Bertrand Russell. Vol. I: 1872-1914. London: George Allen & Unwin. Referred to: '1-'ii. 1968. The Autobiography of Bertrand Russell. Vol. 2: 1914-1944. London: George Allen & Unwin. Referred to: viiin., lxn., xviii, xxxv, xxxvin. 1973. Essays in Ana{vsis. Edited by Douglas Lackey. London: George Allen & Unwin. 1983-. The Collected Papers of Bertrand Russell. London: George Allen & Unwin. 1983. Cambridge Essays, /888-99. Edited by Kenneth Blackwell, Andrew Brink, Nicholas Griffin, Richard A. Rempel and John G. Slater. Vol. I of his 1983-. SCHlLPI', PAUL ARTHGR, ed., 1944. The Philosophy of Bertrand Russell. (The Library of Living Philosophers, Vol. 5.) Evanston and Chicago: Northwestern University. SEARLE, JOHN R., 1957-58. "Russell's Objections to Frege's Theory of Sense and Reference". Analysis, 18: 137-43. SHAKESPEARE, WILLIAM. Macbeth. Referred to: 59. SOMMERVILLE, S.T., 1979. "Types, Categories and Significance". Unpublished PH.n. thesis, McMaster University. Referred to: ixn. 1980. "Wittgenstein to Russell (July, 1913). 'I Am Very Sorry to Hear ... My Objection Paralyses You'''. In Language, Logic and Philosophy: Proceedings of the Fourth International Wittgenstein Symposium. Edited by Rudolf Haller and Wolfgang Grassl. (Schriftenreihe der WittgensteinGesellschafl.) Vienna: HOIder-Pichler-Tempsky. Referred to: ixxn. STOlJT, GEORGE FREDERICK, 1901. A Manual of Psychology. 2nd ed. (The
BIBLIOGRAPHICAL INDEX
anslation '22. ler; New
Marsh.
in. London:
London:
'ge Allen Allen & Andrew . 1 of his
'1I. (The Chicago: ense and
ublished
[ear ... ilosophy: jited by ~enstein-
d. (The
187
University Tutorial Series.) London: W.E. Clive, University Tutorial Press. (Russell's library: 3rd ed., 1913.) Referred to: 2In., 55, 59. 1911. "The Object of Thought and Real Being". Proceedings of the Aristotelian SocIety, n.s. II: 187-205. Referred to: 152n. THOMPSON, 1\1JCHAEL 1975. "Some Letters of Bertrand Russell to Herbert Wildon Carr". Comnto, 10: 7-19. Referred to: xxxv. THORPE, JA.\tES, 1972. Principles of Textual Criticism. San .,>larino, Calif.: The Huntington Library. TWAIN, MARK [pseud.J, 1969. Mark Twain's Mysterious Stranger lvlanuscripts. Edited by William M. Gibson. Berkeley and Los Angeles: University of California Press. UNWIN, PHILIP, 1972. The Publishing Unwins. London: Heinemann. UR.\lSON, j.O., 1956. Philosophical Analysis: lts Development between the Two World iVars. Oxford: at the Clarendon Press. Referred to: viiin. W.-\Rl'IOCK, G.J., 1958. English Philosophy Since 1900. (The Home University Library of Modern Knowledge, No. 234.) London: Oxford University Press. Referred to: xiiin. WHITEHEAD, ALFRED NORTH, AND BERTRAND Rt;SSELL, 1910-13. Principia ,Uathematica. 3 vols. Cambridge: at the University Press. 2nd ed., 1925-27. Referred to: vii, viii, ix, x, xi, xii, xiii, xiv, xv, xviin., xxxvii, xliii, 13n. WIEl'IER, NORBERT, 1914. "A Contribution to the Theory of Relative Position" . Proceedings of the Cambridge Philosophical Society, 17: 441-9. (Russell's library.) Reprinted in Vol. I of his Collected Works: Wich Commentaries. Edited by P. Masani. Cambridge, Mass., and London: The MIT Press, 1976. Referred to: 77n. WITTGENSTEIN, LUDWIG, 1913. "Notes on Logic, September 1913". Translated [as version 1] and rearranged [as version 2J by B. Russell. Manuscript and typescript of version 1 in RAj published in Wittgenstein 1979, App. 1. Version 21st published in The Journal of Philosophy, S4 (1957): 231--45; reprinted in 1st ed. (1961) of his 1979, App. I. Referred to: xxi, xxiii, xxxvii. 1921. "Logisch-Philosophische Abhandlung". Introduction by B. Russell. Annalen der Naturphilosophie, 14: 185-262. Translated into English as his 1922. Referred to: xxxv, xxxvii. 1922. Tractatus Logico-Philosophicus. Translated by C.K. Ogden. Introduction by E. Russell. (International Library of Psychology, Philosophy and Scientific Method.) London: Kegan Paul, Trench, Trubner.
188
BIBLIOGRAPHICAL INDEX
(Russell's library.) Referred to: xxxvi-xxxvii. 1971. Prolractalus: A.n Early Version of Tractatus Logico-Philosophicus. Edited by B.F. McGuinness, T. Nyberg and G.H. von Wright. Translated by D.F. Pears and B.F. McGuinness. Oxford: Basil Blackwell. Referred to: XXX~lll. 1974. Lellers to Russell, Ke)l1les and Moore. Edited by G.H. von Wright and B.F. McGuinness. Oxford: Basil Blackwell. Referred to: ixn., xixn., xxn., xxin. 1979. Notebooks 1914-1916. 2nd ed. Edited by G.H. von Wright and G.E.l.1. Anscombe. Translated by G.E.M. Anscombe. Oxford: Basil Blackwell. 1st ed., 1961. Referred [0: xxin. WOOD, ALAN, 1957. Berlrand Russell; The Passionate Sceptic. London: George Allen & Unwin. Referred to: xxiiin.
hicus . .slated
1t
General Index
and
.E.M. II. 1st
,eorge
PAGE NUMBERS IN roman type refer to Russell's text; page numbers in italics to editorial matter. Editorial matter has been indexed only for proper names and philosophical subjects. For authors, see also the Bibliographical Index.
abstract beliefs 157 factes) 25, 33, 38, 42, 44 ideas: Berkeley and Hume on 91, 95 objects 57, 84-5, 89,131-2 thought 55, 57,129-30 acquaintance with abstract objects 84-5, 89; with complexes 82-4, 120--2, 154, 177; with facts xxxiii; with logical form 99, 100, 101,113,129-31; with subject 36-7; with universals 5, 79, 95-{i, 100, 101 analysis of 5,44,45, 100 definition of 35 and introspection 121 James on 5, 55, 150 kinds of 5, 38, 79 Mach on 5 and memory 71-2,172-3 nature of, summarized 99--101 and proper names 7··8,37,39-40 reality of its objects 48-9 ri val theories of 5 and self-evidence 161, 164 and understanding 108, 130-1 sec also experience; presence act (of presentation) 42, 43 action vs. contemplation 163 affirmation 108 Meinong on 107 Allen (George) & Unwin Ltd. x ambiguity 6, 21-2, 46, 69-70 analysis and attention 123-4, 125-8 complete 119, 120 definition of 119 when insufficient 27 material vs. formal1l9
method in 33 and philosophy 146, 159 and possibifity 27, III problems of 120, 128 and self-evidence 161 and symbolism 146 theory of 125 and time 64, 66 see also construction Annahme: Meinong on 107-8 apparent variable 71, 129, 147 a priori Kant on 22 odium 01'6 Aristotelian Society xii arithmetic II, 13, 30 see also number(s) association of ideas 15 assumptions: Meinong on 107, 108 astronomy 44, 106 atomic thought, summarized 176-8 attention and abstract objects 132 and acquaintance 39-40, 79, 121 and analysis 123-4, 125-8 of eccentric persons 133 and experience 8-9 and logical form 129-30, 132-3 object of 69 as relation 40, 100 awareness 7-8, 35, 72 axiom of time 76n. "before" and "after" 85-8 being: see existence belief abstract 25, 157 analysis of 23-5,46, 142, 144 189
190
GENERAL INDEX
certainty of 167 definition of 177, 178 and doubt 54, 142-3 and dreams 61 emotion of 141 and experience 9 false: see error Hume on 136-40, 177 and ideas 137~40, I SO and inference 47, \OS, 178 James on 32, 140 and judgment 136 and knowledge 156, 160, 178 and logical form \09 Meinong on 107-8, 160 and perception of fact 133 permutative/non-permutative 144-~8, 154 and presentation 141 and self-evideoce ISS, 162, 163~ and sensation 23-5 svmbolized 144 t~ue and false 54 and understanding 108, 137, 142 Bergson, Henri xxvi, 75 Berkeley, George 91, 95 Broad, CD. viin., xxixn. Cambridge University vii, xi-xii, xxi, XXIii Cambridge University Moral Sciences Club xxi Carr, H. Wildon XXX1' Carus, Paul xxv-xx'!}i; causality
in distinction between sensation and imagination 55, 58,60 james on 18, 23, 2&~7 in knowledge of particulars 14 in neurral monism 15 in pragmatism 151 certainty definition of 168 and judgment 176 and logic 168 and memory 168-76 reasons for 163 and self-evidence 167 change (alteration) 67,76,77, 122 see also persistence; recurrence coherence theory of truth 149-50 complex(es) acquaintance with 82~, 120-2, 154, 177 analysis of 83-4, 119~20, 122-3, 146-7, 177 associated 145, 147-8, 154, 178 atomic 80-1, 90,110,120,132,145, 176,177 and change 122 corresponding 144, 14&, 161-2, 165-6
definition of 79-80 dual 80, 82-3, 86-8, 98, 110-16, 131 and facts 79-80 heterogeneous/homogeneous 123, 135, 14-1-7, 178 how they differ 54, 80 and language 87, 114, 145, 146, 147 logical form of 81,87,98-9,113-16, 131,135,146-7,176 molecular 80,135, 145, 147 "the past" as constituent of 70-1 perception of 125-8,154-5,177 permutativelnon-permurative 145~-8, 154 position in 81,88, 92, 100, 111, 113, 122, 145-8 and predicates 90 and propositions 80, III and self~evidence 161-2 subject~predicate 113, 114 symbolized 123, 146~7 symmetrical/unsymmetrical 112, 122-3, 177,178 and synthesis 119 and truth 144-~5 understanding-complex 117-18, 129 conception 10 1, 177 James on 150 consciousness: James on 17-19 construction (logical) 39, 43, 61, 73-7, 122,148 contemplation: see action contents (mental) 5, 41-4 James on 18,23 Meinong on 41~ Perry on 30 contradiction, law of 162 conviction: l'vleinong on 107, 168 Cook, Miss xxv-xxvi, xX1,iin. correspondence (and truth) 25, 109, 148-9, 151~, 162, 164, 165-6, 177 Costelloe, Karin x, xxiii Darwin, Charles 163 data 47,53 deduction 50 definition(s) and science 58 of terms and things: acquaintance 35; analysis 119; anterior and posterior instants 7S, 77; atomic complex 80-1; atomic propositions 105-6, 110; beginning of an event 77; belief 177, 178; certainty 168; cognitive fact 45-6; complete analysis 120; complex 79-~80; complex perception 177; conception 101; concepts 101; corresponding complex 144; data 47; degrees of abstractness 132; dual
GENERAL INDEX
5,
, 154
~-3,
'7
i; ,r }-I;
17, )Iex
47;
relations 80; dualism 7; earlier (and later) 56, 65, 73-4; ego 36; epistemological premisses 50; epistemology 46; existence in time 76; fact 9, 79-80; falsehood 177-8; the future 10, 65, 74; heterogeneous and homogeneous complexes 123; "I" 8, 36, 37, 40, 44; idealistic monism 7; imagination 55-8; immediate and remote past 169; immediate memory 65,70-3; instant 74-5, 76,77; judgment 136; judgmentacquaintance-memory 17l, 172; judgment-descriptive memory 171, 172; knowledge 156; logical form 113-14; logical intuition 101; mental facts 35, 37; mental time 64; mind 35; molecular propositions 105-6; neutral monism 7,15; "now" 8, 29, 65, 69-70; objecl of acquaintance 5, 35, 45; one momentary total experience 6~8; particulars 55-6; the past 65, 71,73,74,77; perception of fact 133; perceptive acquaintance-memory 171, 173; perceptive descriptive memory 171,172-3; permurative/nonpermuralive belief 144-5; persistence in time 76; physical facts 35; physical time 64; precedence (in time) 77; predicates 80-1; prejudice 47; presence (to subject) 64, 68; present experience 8, 37-8; present time 38, 39,65,70,79; proposition 80n., 105-7, 114-16, 134, 177; propositional thought 110; relating relation xx, 80; relations 80, 84; remote past 169; self-evidence 15&-7, 1612, 163, 164, 165-6, 178; sensation 55-8, 63-6; sense-datum 66; simple perception 177; simple terms 120; simultaneity 64-5, 69; specious present 6~8, 72; subject (of experience) 5, 8,35; succession 65, 73; symmetrical wmplex 122; svnthesis 119; "technical philosophv" x-xi; "this" 8, 29, 40; time 66; truth 1445,148,150,154,164,165, 177-8; understanding 108, 110, 17~7; universals 81 delirium tremens 163 Descartes 6, 21 descri ption( s) and "I" 36 knowledge by 11-12, 28, 34, 42, 57, 69, 71,74,81,119,171 and names 3~7, 48,53,94-5, 138,147, 152 and subject (of experience) 35
19]
and universals 29-·30, 81 desire 26 Donnelly, Lucy xxi doubt 54, 142-3., 162--3 dreams 48-9,60-2,63, 150 dual relations acquaintance as a 5, 100 definition of 80 and error 49 and logical objects 97 .'l.ieinong on 108 and predicates 90 and propositions 110, 112, 130-1, 133 and truth 145 understanding of 98 dualism 6-7, 15,21-2,141-2 James on 7, 18 Mach on 16 earlier (and later) 56, 64, 65, 73-4, 77, Ill, 135, 145, 148 ego 36 Mach on 16 see also subject (of experience) elegance (in logic) 50 Eliot, TS. x, xxiii emotion of belief 141 emphatic particulars 40-1 empiricism 11,93,158,159 see also Berkeley; H ume epistemological premisses 50, 51,158, 159 epistemology 46, 50-2, 156 error 23-5, 29,49,62,94,141,154,176 James on 32 Montague on 24-·5 event 77, 107, 114, 115 excluded middle, law of 162 existence 76, 138,141, 151-2, 177 Hume on 139 experience and acquaintance 35 ambiguity of "experience" 5-6, 21 analysis of 5, 33, 46, 64 and attention 8-9 "being experienced together" 8, 66, 79, 99 criterion of 54,73, 142 of experiencing 33-5, 38-9 of fans 9 James on 17-18,23 logical 97 and Iuemorv 73 objects of i-& one momentary total 6~8, 99 of the past 9-10, 33 and peripheral sensations 8-9 present 7-8, 10-12, 14,34,37-40,99 as a relation 35
192
GENERAL INDEX
six questions ~'Oncerning 8 theories of 5 and time 64 transcendanee of 10-12,13-14 unity of 8, 12-13,28-31 see also acquaintance; contents (mental); subject (of experience) extension (of a word) 94 external relations 43, 54 James on 18-19 external world 49,50-1,58,59 factCs) abstract 25, 33, 38, 44 acquaintance with xxv cognitive/non-cognitive 22, 45--6 and complexes 79-80 corresponding 144 and data 47 definition of 9, 79-80 expressed in language 107 logical 5, 30, 33, 47 mathematical 13, 25, 29, 30, 33, 57 mental 22, 35, 37,40,109,131,137, 141, 150 perception of 47, 13 3 physical 35 "primitive" 9 and self-evidence 163, 165 simple 129 and truth 153, 164 form grammatical 48, 93 logical: see under logical F rege, Gottlo b 85 functionC s) mathematical 13 the future 10,65,74 Galton, Francis 61 geometry 22 God: belief in 24 golden mountain 41, 141 grammar 48, 93,106.,139,141,152 see also propositions, sentences; statements gravitation, law of 15 hallucination 24, 48-9, 59,61-3 see also illusions of sense Harvard University x, xi'l1-xv, xx-xxi, xxiii-xx'vi
Hegelianism xiv heterogeneous/homogeneous complexes 123, 135, 144--7, 178 Hume, David on abstract ideas 91,95 on belief 13&40, 177
on on on on
existence 139 the missing shade 89 the self 36 sensation and imagination 53
"I" 8, 36, 37,40,44, 100 see also ego ideals) association of 15 and belief 137-40, 150 Berkeley on 91, 95 doctrine of 22, 26, 139-40 Hume on 91, 95,137-8,139 James on 26 and knowledge of external world 22 and mental wntents 43 Platonic 92 idealism 5--6, 20-1, 22, 24,136,154,158 identity numerical 76, 91,92 personal 12-13 illusions of sense 25., 44 see also hallucination imagination definition of 55--8 and dreams 60-2 and error 62 Galton on 61 Humeon 53 J ames on 150 and memory 9, 57,71-2,170-4 and mental contents 43 and reality 48-9, 53, 59--60,62 as relation 100 and sense-data 176 Stout on 59 and will 58-9 impartiality 40 incomplete symbols 42, 108-9, 113, 114 indubitability 162-3 induction 14, 158-9 inference and belief 47, 105, 178 in BR's method 52 and epistemological premisses 50 and external world 44 and the future 10, 74 in hierarchv of relations 132 and inductive philosophy 154 and matter 21 and memory 174 philosophical views of 158 and reality 49 and self-evidence 156, 157 and sensation 51, 55 and "thing" 93-4 and truth 164 unconscious 61
GENERAL INDEX inspection 54, 66,131,142,153,160,162, 164,178 instant (of time) 74-·7 intension (of a word) 94 introspection and acquaintance 121 certainty of 18 Hume on 36 James on 18-19 and memory 656, 70,72,73 and mental contents 43-4 and mental objects 5, 6-7 and perception 37 and sensation 65~ intuition (logical) 101
.58
14
James, William xv, xvi, xx, xxxz'i, 22 on acquaintance 5, 55, 150 on belief 32, 140 on causalitv 18,23,26-7 on consciousness 17-9 on dualism 7, 18 his idealism 20-1 his neutral monism 7, 14, 138, 150 his "processes of leading" 26-8 on knowledge 17,19-20,26-8 on the past 72 Perry on 21 on specious present 72 Joachim, H.H. 149 Jourdain, Philip ix, xxii-xxiii, xxvii, xxix judgment analytic 162 and certainty 176 and data 47 definition of 136 existential 138, 151·2 Hume on 137-8 and logic 46, 136 logical structure of 110 Meinong on 41, 108 and memory 171-4 of sense 167 and truth 79, 108 see also belief Kant, Immanuel 19, 22 Keynes, J.M. xxixn. knowledge: difficult to define 156 language and certainty 168 and comple;es 87, 114, 145, 146, 147 European 107 and facts 107 and relations 85-8 symbolic 93 see also ambiguity; description(s);
193
grammar; "I"; meaning; names; "now"; statement; "this" Lausanne xvi, xvii Leibniz, G.W. von 6 Lenzen, Victor x, xXli, xxiii-xxv logic and certainty 168 and deduction 50 and epistemology 46, 50 and excluded middle 162 and idealism 136 and judgment 46, 136 kinds of 149 and predicates 90 of relations 13n. traditional 94, 105 sec also inference; probability logical constant 98 construction 39, 43, 61, 73-7,122,148 experience 97 fact 5, 30, 33,47 form: and abstract objects 131; acquaintance with 99, 100, 101, 113, 129-31,141; and attention 129-30, 132-3; and belief 109; of complex 81, 87,989,110-16,131,135,146-7, 176; definition of 113-14; and falsehood 141; of memory 173; of mental fact 131; molecular 132; and propositions 46, 110, 129-30; of propositions 11,86,93,94,98, Ill, 129; and understanding 108, 129-31, 137 intuition IOJ judgments 168 objects: not entities 97 premisses SO, 158-9 propositions 97-8, 132, 177 structure 113, 116 Lowell Lectures xiv-xv, xx-xxi, xxiii, xxv,
xxviii, xxxv Mach, Ernst xZ'i, 5, 16, 17 Perry on 21 Marsh, R. C. x mathematics II, 13, 22 see also arithmetic; see also under fact(s) matter 22 Descartes on 21 meaning 36, 106, 109, 110, 134 Meinong, Alexius xiv, XXi'll} xxxii, xxxiii on belief 107-8, 160 on contents (mental; 41-4 on memory 168 ..9 on presumptions 169-70, 176 on propositions 108, 110 on self-evidence 160-1, 163-4
GENERAL INDEX
194 on truth 152-3 memory
and acquaintance 71-2,172-3 analysis of 56,57,64,70-3 and certainty 168-76 and error 49 expressed as logical relation 13n. and ideas 26 and imagination 9, 57, 71-2,170--4 immediate 65, 70-3,169,173 intellectual 9-10 and introspecrion 65-6, 70, 72, 73 James on 150 judgment-acquaintance-memory 171, 172, 174 judgment-descriptive memory 171, 172, 174 Mcinong on \68-9 perceptive-acquaintance-memory 171, 173,174 perceptive-descriptive memory 171, 172·3, 174, 175 and personality 12 physiological 72 sensational 9--10 and solipsism II and unity of experience 12-13 metaphysics 5, 64,75,93, 157 mind(s) 6, 13,35,42-3 Perry on 30 monism
idealistic 6, 7, IS, 154, 158 materialistic 15 neutral: see neutral monism The Monist xxii ,\\ontague, W.P. 24-5 Moral Sciences Club: see Cambridge University Moral Sciences Club Morley, Percy Hz,;i lviorrell, Ladv Ottoline vi., ix, xii, xvi, XVH
names complex 128, 148, 177 and descriptions 3&-7, 48, 53, 94-5, 138, 147, 152 proper: see proper names negation 106, 108, 199 Meinong on 107 negative propositions 152 nervous system 51 Perry on 30 Stout on 55 neutral monism definition of 7, 15 and emphatic particulars 40-1 grounds for 21-2, 36 and James 7,14,138, ISO
objections to 23-32, 33,40-1,43,44,93 its service to philosophy 31, 34 and Stout 21n. set' also James; Mach; ,\lontague; PelTv "now" 8, 29, 65, 69-70, 100 numbers II numerical identity 76, 91, 92 ObjekllZ'; Meinong on 108, 110, 160 Occam's razor 21 Ogden, C.K. xxiii Open Court Publishing Company x, xxii
xxv-·xxvii other minds 13-14 particular(s) and abstract objects 131 and acquaintance 5, 100 and attention 129 beyond experience 14 and common-sense "thing" 91-2 definition of 55--6 and data 47 emphatic 4!f I and facts 35 and imagination 56, 58 and sensation 55--6, 58 and universals 93, 94-5 see also "1"; ilnow"~ "this" Parmenides 49 the past 9-10, 56, 65,70--4,77,169-71 James on 72 patriotism, emotion of 141 perception 37,123, 125-8,133,154-5, 177 James on 55 permutation-groups 123 permutativeinon-permmative beliefs 144-8, 154 Perry, R.B. x, xx, xxii, 21,29,30 persistence ,in time) 76,78 personal identity 12-13 philosophers beliefs they should a\'Oid HI cunning 6 impatient 5 misled by grammar 139 who dislike analysis 122 philosophy Cartesian 6 confusion in 93-5, 152 and detachment 159 empirical: see empiricism inductive 154 method in 6, 21, 33, 55, 155 technical x-xi of time 70, 107 the world considered without 7
GENERAL INDEX see also dualism; idealism; metaphysIcs; monism; ft:alism; scepticism
physics 15,21,22,50-2,54,57,60,62 Mach On 16, 17 see also aSlfOnt)my~ gravitation; psychophysics physiologv 31,502, 55,58,60,72, 9S Plato xi, xiv, 92 political belief 141 position (in a complex) 81, 88, 92, 100, 111,113,122,145-8 possibility 27, Ill, 122, 152 pragmatism 26,149,150--1,154 precedence (in time) 77, 81 predicate(s) an analyticity 94-5 and common-sense "thing" 91-2, 9lA and complexes 90 definition of 80-1 reality of 90 I, 92-3 and self-evidence 163-4 and sense-data 91-2, 94, 95 truth as a 164 and universals 81, 90, 92 prejudice 47 premiss(es) 50, 51,158, IS9 presence (to subject) 64, 68 the present (time) 10, 38, 39, 65, 70, 74, 79; see also specious present presentation 141, 152 Meinong on 41-3,107,108 presumption(s): Meinong on 169-70, 176 primary constituents 120 prime numbers II prison (of experience) 10 privacy (of experience) 3+-5 see also solipsism probability 158-9, 168, 175, 176 l,\einong on 169 proper names and descriptions 27, 36-7, 48, 138, 147, 152 and logical form 130 objects of 7-8, 39-40, 44, 65 proposition( s; analysis of 139, 141 atomic 105-6, 110, 1:6, 129-30, 135 and complexes 80, III and correspondence theory of truth 149 definitionof80n., 105-7, 11+-16,134, 177 epistemologIcally self-contained 50 existential 138, 140, 152 logical 97-8, 132, 177 and logical form 46, 110, 129-30 logical form of 11, 86, 93, 94, 98, Ill, 129 Meinong on 108, 110
195
molecular 105~6, 116, 130, 139, 145-6, 148, 153, 154, 162 Montague on 24.. 5 negative 152 non-mental 153, 155 and fealily 25, 109-10, 113 and self-evidence 157 synthetic 95 and truth 108-10, 144 understanding of 176-7 and universals 133, 177 propositional synthesis 119 propositional thought, kinds of 110 as men tal relation 37 as physical relation 37 psychology and belief 136 and deduction 50 and epistemology 46 and imagination 62 and James 17 and logical form 132 on memory 72-3,173·4 and space 22 psychophysics 65 rationalism 150 realism 6, 7,92 reality and correspondence theory of truth 149, 151-2 and imagination 48·9, 53, 59-{)0, 62 Montague on 2+--5 of objects of acquaintance 48-9 of predicates 90-1,92·3 of propositions 25, 109-10, 113 Stout on 152n. transcending present experience 10-12 reason 163 recurrence (in time) 74-6, 78 relating relation acquaintance with 82 analvsis of 83, 120, 146-7 in complexes 90,123,139,176 definition of xx, 80 in a fact 131 and position in complex 88, 100, III and propositions 115 sense of 86-9, 100, 111-12, 116, 134-5, 145, 147 and understanding 117 relation( S) classified 80 cognitive, hierarchy of 131·2 and complexes 54 definition of 80, 84 dual: see dual relations external 43, 54
GENERAL IKDEX
196
form of 86 hierarchy of 131-2 internal 43, 44 James on 18, 19 logic of 13n. mUltiple 115, 131, 137, 153, 177 and Platonic ideas 92 and self-evidence 162 serial 77 simple (or ultimate) 66 Stout on 21n. symmetrical/unsymmetrical: and predicates 91-1; and sense of a relation 86, 145, 147; and similarity 117; and temporal relations 65, 74, 75,77; and truth 149; and understanding 133, 134-5 terms of 86--7, 90, 98 three-term 133 and time 64, 66, 76, l70-1 transitive 65, 74-5, 77, 78, 91, 92, 149 two-term 53, 70, 75, 131 as universals 81 religious belief 141 representation 176 resemblance 82-4 Robinson, Lydia xxv, xxviin. round square 41, 42,141 Russell, Bertrand collaboration with Whitehead vii, xiii, xxi commentary on his philosophy viiin., xii-xiii his Harvard appointment xiv-xv, xx-xxi, xxiii, xxvi, xxviii his philosophical mission in America xiv his philosophical projects, 1910-13 xii-xiv his philosophical reputation viin. interaction with Witrgenstein vii, ix, xi--xii, x'vii-xxii, xxviii, xxxiii-xxxvii Lowell lectures xh~xv, xx-xxi, xxiii, xxv, xxviii, xxxv
opposition to World War I vii, xxxv-xxxvi relationship with Ottoline Morrell vii, xi, xvii, xix, xxv on "technical philosophy" x-xi Sanger, C.P. xx; scepticism 51,156,159,162-3 science 21,31,33, 158 see also physics secondary constituents 120 self: see subject self-consciousness 38 self-evidence and belief ISS, 162, 163-4
and certainty 167 definition of 156--7, 161-2, 163, 164, 16S--{),178 and premisses 159 Memong on 160--1, 163-4 sensation(s)
and belief 23-5 and causality 55, 58 definition of 55-8, 63--{) and dreams 50--2 as a dual relation 100 "force or liveliness" of 59 Hume on 53 James on 19, 55, 150 logical structure of 110 Mach on 16 and perception 37 peripheral 8~9 reality of 53, 59--{)O, 62 and scepticism 51 and self-evidence 157 Stout on 21n., 55, 59 and time 38, 64, 66, 77-8, 79 and will 58-9 sense (of a relation) 86--9, 100, 111-12, 116,134-5,145,147 sense-data definition of 66 divisibility of 121-2 and imagination 171, 172, 173, 176 and memory 72, 73, 171-3 and "now" 69~70 and physics 53-4 and predicates 91-2, 94, 95 and specious present 68 and succession 73 and things 93-4 sensible continuum 124 sentences 105-6, 144 sequence (in nme) 88, III, 134, 135, 145 similarity 112, 116--18, 133 simultaneity 64-5,69-70,74-7,79, 161, 165 Socrates 49 solipsism 10-12, 13-14,34 space 21-2,91,92,93,121-2 specific similarities 91 specious present 66--8, 69--70 James on 72 Spinoza, Benedictus de xi statements 105-6, \O8~9 Stephen, Karin: see Costelloe, Karin stimulus 55, 58, 59, 60, 72 Stout on 55, 59 Stout, G.F. 2In., 55, 59, 152n. Strachey, Oliver ix, xx subject (of experience) 5, 8, 35-9, 44-5, 170-1
1
GENERAL INDEX Hume on 36 Mach on 16
see alsa "I" subsistence 42,91,153 Meinong on 41, 152 substance 15, 22, 93 James on 17~~18 sllccession 64-5, 70, 71, 73, 75, 77-8, 79 symbolic language 93 symbols 42, 108-9, 113, 114, 144, 146 symmetricaVunsymmetrical: see under complex(es); relatione,) synthesis 101, 119, 128 synthetic: see under proposition(s) terms of complexes 80, 83-4, 88, 100, 116-18, 120, 126-7 in logic 105 of relations 86-7, 90, 98 thing 91~2, 93-4, 98 "this" definition of 8,29,40 and experience 8,10,67,100 and "I" 40, 44 and "now" 69~70 and the past 71 as proper name 39-40, 48, 65, 95, 138 thought 30, 55, 57, 110, 129~30, 147, 176-8 time absolute 69, 70, 74, 76 and abstract thought 57 and acquaintance 5, 81 axiom of 76n. "before" and "after" 85~8 Bergson on 75 BR's lecture on xxin~xxiin. definition of 66 divisibilitv of 121-2 and lan~age 107 and metaphysics 93 mental 64 and mind 35 physica164 precedence in 77, 8 I recurrence in 74-6, 78 see alsa change; earlier (and later); the past: the present; sequence; simultaneity Trinity College, Cambridge xxiIi, xxxv truth analysis of 154 coherence theorv of 149-50 correspondence 'theory of 25, 109, 148-9,151-4,162,165,177 definition of 144-5, 148, 150, 154, 164, 165, 177-8
2,
6
;,145 161,
n
t4-5,
1
197
and fact 153, 164 and falsehood 46, 141-2, 153, 177-8 Hume on 139 James on 26, ISO Joachim on 149 Meinong on IS2-3 Montague on 24 necessarv 134 and pragmatism 26,149, ISO-I, 154 as a predicate 164 and propositions 108-10, 144 and self-evidence 157, 161, 163, 164 theory of 147,148,151-2 and verification 27 understanding and acquaintance 108, 130-1 analysis of 131 and belief 108, 137, 142 definition of 108, 1l0, 176-7 and logical form 108, 129~31, 137 of a proposition 176-7, 195 understanding-complex 117-18, 129 and universals 133 universal( s) and abstract objects 131 acquaintance with 5, 79, 95~-6, 100, 101 and attention 129 and certainty 168 and data 47 definition of 81 and descriptions 29-30, 81 and ego 36 knowledge of 44 and particulars 93, 94-5 and predicates 81, 90, 92 and propositions 133, 177 and understanding 133 unreality 60, 62, 63, 109, 131, 151~2 vagueness 174-1l variable(s) 13,37,71,98,113,129,147 verification 27 Vittoz, R. xvi, xviii Watson, J .B~ xxxv. Whitehead, A.K. vii, xiii, xxi, xxxv Wiener, Norbert 77n. wi1158-9,60 Wittgenstein, L. 46n. acknowledged by BR xxix, xxxii, xxxiii, XXXIV, XXXV!
his criticism of Theory of Knowledge ix, xvii-xx, xxii, xxvii his work on logic xxi, xxxvi influence on BR vii, xi-xii, xxx'l>-xxxvii personal characteristics xi world 49, 50-1,58,59 ,1,1.ach on 16